Abstract
The aim of the paper is to investigate the effect of size of multiwalled carbon nanotubes (MWCNTs) as additives for dispersion in gear oil to improve the tribological properties. Since long pristine MWCNTs tend to form clusters compromising dispersion stability, they are mildly processed in a ball mill to shorten the length and stabilized with a surfactant before dispersing in lubricant. Investigations are made to assess the effect of ball milling on the size and structure of MWCNTs using electron microscopy and Raman spectroscopy. The long and shortened MWCNTs are dispersed in EP 140 gear oil in 0.5% weight. The stability of the dispersed multiwalled carbon nanotubes is evaluated using light scattering techniques. The antiwear, antifriction, and extreme pressure properties of test oils are evaluated on a four-ball wear tester. It is found that ball milling of MWCNTs has a strong effect on the stability and tribological properties of the lubricant. From Raman spectroscopy, it is found that ball milling time of up to 10 hours did not produce any defects on the surface of MWCNTs. The stability of the lubricant and the antiwear, antifriction, and extreme pressure properties have improved significantly with dispersion shortened MWCNTs. Ball milling for longer periods produces defects on the surface of MWCNTs reducing their advantage as oil additives.
1. Introduction
Gear oils used in industries and automotive engines are often subjected to heavy loads, due to which they experience high temperatures and pressures causing higher friction and surface damage leading to failure of the system. To prevent failure, conventional engine and gear oils are dispersed with extreme pressure (EP) and antiwear (AW) additives that react chemically with the metal surfaces, forming easily sheared layers and thereby preventing severe wear and seizure. Allotropes of carbon such as graphite, fullerenes, carbon nanotubes, and graphene have attracted the interest of the researchers due to their special properties. The hybridization of the atomic orbital of carbons in carbon nanotubes and fullerenes is of type sp2, similar to that of graphite, making a perfect hexagonal array of atoms. The sp2 C–C bond of CNTs is considered one of the strongest in solid materials; thereby CNTs are expected to yield exceptionally good mechanical properties. Lubricants dispersed with allotropes of carbon are being extensively studied for their lower friction coefficients, thereby improving antiwear properties. MWCNTs possess large surface area compared to many inorganic nanomaterials and can be very easily surface-modified. Several novel studies are made on the effect of dispersion of multiwalled carbon nanotubes on the wear and friction characteristics of lubricants. The length of the nanotube synthesized by existing methods is known to be thousands of times larger than their width and thus limits their functionality for many applications. Owing to their large length to diameter ratios, MWCNTs despite use of surfactant tend to form agglomerates faster, thereby leading to them settling in the liquid medium. This aspect is a main challenge in obtaining stable dispersion in liquid medium. The most common and frugal method to reduce the size of MWCNTs is ball milling which shortens the length of MWCNTs and obtaining open ends. However, ball milling for extended period produces defects and damages the graphite structure rendering it useless for dispersion in lubricants. Studies are made on the effect of ball milling on the structure and defect generation. Pierard et. al. [1] studied the effect of ball milling on the structure of single-walled carbon nanotubes. Raman spectroscopy was employed to study the defects produced in various hours of ball milling time. It is found that there exists an optimum time to keep the tubular structure intact without defects. In case of single-walled carbon nanotubes, ball milling times of over 50 hours completely destroy the structure producing amorphous carbon. Dresselhaus et. al.[2], Cancado et. al. [3], and Paton et. al. [4] suggested methods to detect defects and evaluate purity of carbon nanotubes and graphene. It was proposed that the intensity of G-band and the D/G ratio can be highly useful to determine both the purity and the defect density of carbon nanotubes and graphene. Chen et. al. [5] first studied the effect of dispersed of ball-milled and stearic acid modified multiwalled carbon nanotubes (MWCNTs) on the stability and thereby improvement in the lubricating properties of liquid paraffin base oil. The friction and wear tests were conducted on a pin-on-plate wear-testing machine. It is found that ball-milled MWCNTs could form stable suspensions and improve the antiwear and antifriction properties. Engine oils dispersed with nanomaterials are investigated for tribological property enhancement by several researchers. A detailed review of the prior art is listed in Table 1.
As given in Table 1, most of the studies were done with dispersion of larger amounts of MWCNTS in base lubricants or paraffinic oils without additives. Only few studies were made on commercial grade formulated lubricating oils. Further in the studies, due to dispersion of MWCNTs in higher concentrations, the viscosity of the lubricant is found to be significantly enhanced. This enhancement in viscosity is the major reason for the improved tribological properties. Furthermore higher concentrations of any nanomaterials may lead to faster agglomeration rates and the stability of the suspension will get compromised. Since MWCNTs have long length ranging from 1 to 25 microns and diameter in nanometers, there is a strong ability of entanglement of individual nanotubes leading to formation of clusters. These clusters tend to become hard and make the MWCNTs lose their special properties. Moreover, the stability of nanofluid has not been assessed quantitatively in the studies made so far. Further the defects produced in MWCNTs due to ball milling and its consequential effects on tribological properties also need to be assessed. The present study is aimed at investigating improvements in antiwear, antifriction, and extreme pressure properties of formulated EP 140 grade gear engine oil dispersed with surface-modified and ball-milled MWCNTs. Ball milling is performed in inert gas medium and at a low intensity to prevent damage to the structure of MWCNTs. A simpler surface modification technique is used to stabilize the MWCNTs in oil medium and the stability of the suspensions is investigated over a period of 2 months. The effect of decrease in length of MWCNTs due to ball milling on the tribological properties has been studied. The repeatability of results over a period of 60 days is investigated to observe consistent performance and the average values are reported. The effect of additives in the oil along with surface-modified MWCNTs has resulted in use of lesser amount of MWCNTs in the lubricant dispersion compared to studies reported in the literature which is one of the novel features of the study. The paper also compares chemical and physical routes for dispersion of MWCNTs in lubricant by comparing performance of surface-modified pristine MWCNTs and surface-modified ball-milled MWCNTs, respectively.
2. Experimental
2.1. Materials
In the present study, multiwalled carbon nanotubes produced by CVD method have been procured from M/s Cheap Tubes Inc., USA. The size of MWCNTs is 20-40 nm in diameter, 25 microns in length, and 95% of purity. All other chemicals purchased are of GR grade. The surfactant is AR grade procured from M/s Sigma Aldrich India Pvt limited. GL4 (EP 140 grade) gear oil is selected as base lubricant.
2.2. Ball Milling of Multiwalled Carbon Nanotubes
As the length of the carbon nanotubes is thousands of times larger than their width, ball milling of MWCNTs is a common procedure to generate short and open-ended nanotubes. Ball milling apparatus consists of tungsten carbide lined bowls containing tungsten carbide lined ball. Arrangement is provided to perform ball milling under argon atmosphere to prevent oxidation of material during ball milling. The speed of rotation of bowls is set at 400 RPM and the ratio of balls to MWCNTs is taken as 10:1 to ensure minimal damage to the tubular structure of the MWCNTs. Higher speeds and ratios will increase the impact and thereby attrition of MWCNTs. Ball milling was performed for 4 and 20 hours at 400 RPM to avoid damage to the structure. After ball milling, the MWCNTs are heated in air at 600°C to remove amorphous carbon generated during ball milling process. These ball-milled MWCNTs are characterized using HRSEM and transmission electron microscopy to determine the average length of the MWCNTs.
2.3. Electron Microscopy
Figure 1(a) shows HRSEM image of pristine long length entangled MWCNTs. Figure 1(b) shows MWCNTs ball-milled for 5 hours with a small change in length of MWCNTs as compared to Figure 1(a). Figures 2(a) and 2(b) show images of 10- and 20-hour ball-milled MWCNTs.
(a)
(b)
(a)
(b)
It can be seen that the average length of 10-hour ball-milled MWCNTs is under 4 microns. From Figure 2(b) (TEM image) it can be observed that the length of 20-hour ball-milled CNTs has come down to around 150 nm size.
2.4. Raman Spectroscopy
Raman spectroscopy is employed to assess the formation of defects during ball millings and shown in Figure 3 for pristine MWCNTS, 5-, 10-, and 20-hour ball-milled MWCNTs. The sp2 structure of MWCNTs causes first order peaks D and G bands that are approximately located at 1350 cm−1 and 1580 cm−1, respectively. Defect free MWCNTs due to intact hexagonal graphite structure make the G-band sharper. Defects in MWCNTs make the G-band peaks wider and shorter. On the other hand, D band peak represents lattice defects and finite crystal size. During defect formation, due to breaking of the 2D translational symmetry the D band peak will increase and become wider. Another peak band which can be seen at 2700 cm−1 Raman Shift represents amorphous defects in MWCNTs.
Table 2, provides the intensities of D, G, and bands. In case of pristine MWCNTs, 5- and 10-hour ball-milled MWCNTs, there is no significant difference in the intensities of D, G, and bands with ratio of intensities of D and G bands remaining marginally the same. In case of 20-hour ball-milled MWCNTs, a significant decrease in intensity of G and bands is observed with increase of intensity of D band indicating mild destruction of graphite structure and formation of amorphous defects. In all cases, the peaks of all bands are sharp indicating either no defects or mild defect (in case of 20-hour ball-milled MWCNTs).
2.5. Surface Modification of MWCNTs
Pristine MWCNTs tend to agglomerate and form large particles clusters in liquid medium. Moreover, after the ball milling process, the MWCNTs tend to be compressed by the balls forming larger aggregates of entangled MWCNTs. To disentangle them and make them stable in the lubricant medium, it is required to modify the surface of MWCNTs with a surfactant to create stearic repulsions between individual nanotubes. To stabilize the nanoparticles in the liquid medium, a surfactant SPAN 80 (Sorbitan monooleate) is used to modify surface of MWCNTs during the preparation of oil. SPAN 80 is a nonionic surfactant with a hydrophilic-lipophilic balance of 4.6 which is ideally suitable for oils. It adsorbs on the surface of MWCNTs reducing their surface energy, thereby preventing agglomeration and settling of nanoparticles. To prepare surface-modified MWCNTs, SPAN 80 and multiwalled carbon nanotubes (both pristine and ball-milled) are taken in the ratio of 2:1 and ultrasonicated in a solvent for 30 minutes, which creates a mechanochemical reaction. This reaction coats the surfactant on to the surface of the MWCNTs.
2.6. FTIR Spectroscopy
The surface-modified MWCNTs are characterized for functional groups on the surface using Fourier transforms infrared spectroscopy as shown in Figures 4 and 5. Figure 4 shows pristine MWCNTs with no characteristic peak detected. Figure 4 shows the FTIR image of surface-modified MWCNTs with characteristics peaks between 1463 and 1486 cm−1 wavelength indicating lipophilic groups attached to the surface.
2.7. Preparation and Evaluation of Stability of Lubricants with MWCNTs
The surface-modified long and ball-milled MWCNTs in 0.5 weight percent are dispersed in lubricating oil by processing it in a probe ultrasonicator (Hielscher UP400S) for about 30 minutes. The stability of the nanofluid for a period of 60 days is monitored by light scattering techniques. Ball-milled MWCNTs when dispersed in lubricating oils could form better stable suspension compared to long MWCNTs and gave better tribological properties.
The stability of the lubricants dispersed with MWCNTs is evaluated using light scattering techniques by means of zeta sizer (Horiba SZ 100). The zeta potential of the samples, an indicator of dispersion stability of MWCNTs in the lubricating oil medium, has been analyzed over a period of 60 days. A zeta potential value of ± 40 indicates good stability. As the viscosity of the gear oil is high, the oil samples are diluted using toluene before charging it into the zeta sizer to improve the transmittance of the oil so that accurate values can be obtained. Higher values of zeta potential values are found when the oil samples are dispersed with ball-milled MWCNTs. The variation of zeta potential immediately after preparation and 60 days after preparation is shown in Figures 6 and 7.
(a)
(b)
(c)
(d)
(a)
(b)
(c)
(d)
Figures 6(a) and 6(b) show the variation of zeta potential of lubricant dispersed with long MWCNTs and 5-hour ball-milled MWCNTs, respectively. As can be seen the zeta potential with long MWCNTs is much less indicating low stability. With five-hour ball-milled MWCNTs there is an improvement in the stability compared to stability of long MWCNTs. Figures 7(a) and 7(b) show the variation of zeta potential of lubricant dispersed with 10-hour and 20-hour ball-milled MWCNTs. As can be seen there is a significant improvement in the stability with 20-hour MWCNTs showing best stability. This can be attributed to lower agglomeration rates due to short length of MWCNTs.
2.8. Physicochemical Properties of Test Oils
The gear oils are manufactured by blending base stocks and additive components to meet the requirements of standards. Basic physicochemical properties should be in compliance with international standards for statutory purposes. The main physicochemical properties to be assessed for a gear oil are viscosity, viscosity index, pour point, flash point, total acid number, and copper strip corrosion resistance. All the properties for test oils are evaluated in 5 replicable experiments and the average values are reported.
As can be observed from Table 3, the effect of ball milling and surface modification has practically no effect on the physicochemical properties of the test oils. The viscosity and viscosity index are unchanged with dispersion of nanomaterials. Surface modification has no effect on pour point, flash point, and total acid number of the test lubricants.
2.9. Tests for Tribological Properties
The test oils are tested for improvement in tribological properties on a four-ball tester. The weight percentage of surface-modified and ball-milled MWCNTs is maintained as 0.5 wt %. The standard code of the tests, a rotating steel ball, is pressed against three steel balls firmly held together under a load and immersed in lubricant. The test parameters of load, duration, temperature, and rotational speed are set in accordance with standard test procedure.
In wear test done as per ASTM D 4172, the average scar diameter formed on the bottom of three balls shows the ability of the lubricant to prevent wear. A larger diameter indicates poor antiwear behavior. The test is carried out for one hour at a load of 40 kgf and speed of 1200 RPM with temperature of oil maintained at 75°C.
The friction test is carried out to find the friction coefficient offered by the lubricant as per ASTM D 5183 code. Initially the balls are subjected to “wear in” for one hour at a load of 40 kgf and speed of 600 RPM with temperature of oil maintained at 75°C. After “wear in”, the used lubricating oil is discarded and balls are cleaned. Fresh lubricant sample is taken in the ball cup with the same worn test balls in place. The test is again started under the above conditions with load varying from an initial load of 10 kgf and increasing by 10 kgf at the end of each successive 10 min interval until there is a sharp rise in the Frictional Torque which indicates incipient seizure. This is called seizure load which is an important factor in determining the effectiveness of the lubricant.
ASTM D 2783 specifies the extreme pressure properties of lubricant in terms of weld load which is the ultimate load at which the lubricant evaporates due to high pressure and temperature resulting in all the four balls welded to each other. The standard also specifies another parameter called “load wear index” which indicates the behavior of lubricant in resisting the aforementioned weld conditions. The test is used to determine the load carrying properties of a lubricant at high test loads usually encountered in gears. In this test on four-ball tester a series of 10 tests of 10-second duration are carried with varying load under the following conditions:
temperature of oil: room temperature,
speed of rotation: 1760 RPM,
duration: 10 s,
load applied: 32 kgf to weld load.
A total of 10 readings are considered in the test and the corrected load is calculated for all ten readings. The load wear index is a single parameter that shows the overall EP behavior in a range between well below seizure and welding is calculated from the corrected load.
where L is the applied load, kgf, X is the average scar diameter on the worn balls in mm and Hertz scar diameter, and in mm. The load wear index is calculated from the expression LWI = (A/10) (kgf), where A is the sum of the corrected loads determined for the ten applied loads immediately preceding the weld load.
3. Results and Analysis
All the tests are conducted in ten repeatable trails over a period of 60 days to ascertain the influence of nanoparticles in terms of repeatable and reproducible results, and the average values are reported. The results of wear test conducted as per ASTM D4172 are as given below for different test oils. From Table 4, it can be observed that the average wear scar of lubricant with MWCNTs is much less than that of base lubricant. Further, with dispersion of long MWCNTs, the range of wear scar over 60 days of testing is higher compared to other oils due to poor lubricant suspension.
Table 5 shows the results of friction test in terms of average friction coefficient and seizure load. With dispersion of ball-milled MWCNTs, there is a good improvement in both seizure load and the friction coefficient.
The 10 sets of results of repeatable friction characteristics are plotted in graphs in Figure 8. In case of lubricant dispersed with long MWCNTs, although the performance improved during the initial days, there is steady decrease in the performance characterized by increase in friction coefficient due to poor stability over a period of time. Shortened MWCNTs due to their better stability gave a consistence performance over a period of 60 days with lubricant dispersed 10-hour ball-milled MWCNTs giving best performance. The variation of friction torque with time is plotted in Figure 9. From the graph it can be seen that the effect of nanomaterials is more significant at higher loads. In boundary lubrication regime there is significant contact between surfaces separated by a thin lubricant film. Increase of normal load would sweep the lubricant out of the contact region, reducing the lubricant film thickness between surfaces and increasing the chance of contact between surfaces in motion. Due to short length, the ball-milled MWNTs dispersed in lubricant could effortlessly glide and roll between the two contact surfaces in motion like spacers. This increases the pressure limits of the lubricant, thereby significantly reducing friction coefficient and improving the seizure load. Lubricant dispersed with ball-milled MWCNTs has shown consistent performance on the torque-time plot compared to base lubricant.
The extreme pressure properties, namely, last nonseizure load, weld load, and load wear index of test oils under consideration, are summarized in Table 6.
With dispersion of MWCNTs, there is an improvement in last nonseizure load, weld load, and load wear index. A graph showing the variation of wear scar diameter with load is shown in Figure 10. The region between points 0 to 1 is normally designated as antiwear region in which all the lubricants exhibited similar behavior. Points 1 and 1′ indicated the last nonseizure load (LNSL) up to which the wear scar formed is uniform.
The region above last nonseizure load is called extreme pressure zone in which the efficacy of additives in the gear oil comes into play. In this region, the pressures and temperature are very high and the additives are supposed to withstand these extremities. MWCNTs due to their good mechanical properties could form a barrier between the surfaces withstanding the extremities to improve weld load. There is a reduction in the wear scar with dispersion of pristine and ball-milled MWCNTs. The effect of dispersion of shortened MWCNTs is found to be quite visible in this region. Lubricant dispersed with 10-hour ball-milled MWCNTs improved the LNSL and could reduce the wear scars even in the extreme pressure regions. Lubricant dispersed with ball-milled MWCNTs due to decrease of the friction coefficient offered lower wear scars on the test balls. Further after last nonseizure load, it can be observed that the wear scar diameter with dispersion of nanomaterials is significantly low, resulting in improvement in load wear index as well as weld load. This is due to better dispersion owing to short length. Here the ball milling timing of MWCNTs plays an important role in defining the optimum length with minimum damage to the tube structure. The optimum length for attaining good stability and tribological properties can be assessed as 1 to 5 microns, which can be attained with 10 hours of ball milling at 400 RPM speed. 20-hour ball-milled MWCNTs could improve the load wear index by performing well in the antiwear region but, due to defect formation during ball milling, exhibited lower performance in EP region. This can be attributed to the formation of defects on the surface (Figure 3 and Table 2) leading to loss of special properties of MWCNTs.
4. Conclusions
(1) The ball milling of multiwalled carbon tubes prior to dispersion in lubricant plays an important role in improvement of stability, antiwear, antifriction, and extreme pressure properties of gear oil.
(2) Ball milling could shorten the MWCNTs making them stable in the lubricant for a period of more than 60 days.
(3) From Raman spectroscopy it can be observed that ball milling time of up to 10 hours did not produce any defects on the surface of MWCNTs but 20-hour ball milling produced mild defects on the surface.
(4) Long MWCNTs, although surface-modified, when dispersed in lubricants exhibited poor stability and could marginally improve the antifriction and extreme pressure properties.
(5) The physicochemical properties remain unaltered with dispersion of surface-modified and ball-milled MWCNTs.
(6) There is a good improvement in the wear scar diameters and friction coefficients with dispersion of shortened MWCNTs as short-length MWCNTs could slide between mating surfaces reducing contact.
(7) The load wear index and weld load of lubricants dispersed with shortened MWCNTs have improved significantly as the performance of short MWCNTs is good in both antiwear and extreme pressure region.
(8) At an optimum length of MWCNTs in the range of 1 to 3 microns, the lubricant could give best results while too short MWCNTs due to formation of defects despite forming good suspension could not give best performance.
Nomenclature
| : | Hertz scar diameter |
| EP: | Extreme pressure |
| AW: | Antiwear |
| L: | Load applied |
| LNSL: | Last nonseizure load |
| LWI: | Load wear index |
| MWCNTs: | Multiwalled carbon nanotubes |
| SL: | Seizure load |
| X: | Average scar diameter. |
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
The authors gratefully acknowledge the support received from Hindustan Petroleum Corporation Ltd., India, for conducting the tests. The authors acknowledge the assistance from IIT Madras, IIT Kharagpur, and Osmania University, Hyderabad, in characterization. The authors sincerely thank the management of GITAM deemed University, India, for the support extended.