Effects of vibration frequency and amplitude on friction reduction and wear characteristics of silicon

Effects of vibration frequency and amplitude on friction reduction and wear characteristics of silicon

Tribology International 94 (2016) 198–206 Contents lists available at ScienceDirect Tribology International journal homepage: www.elsevier.com/locat...

6MB Sizes 29 Downloads 77 Views

Tribology International 94 (2016) 198–206

Contents lists available at ScienceDirect

Tribology International journal homepage: www.elsevier.com/locate/triboint

Effects of vibration frequency and amplitude on friction reduction and wear characteristics of silicon Shin-Sung Yoo, Dae-Eun Kim n School of Mechanical Engineering, Yonsei University, 50 Yonsei-ro, Seodaemun-gu, Seoul 120-749, Republic of Korea

art ic l e i nf o

a b s t r a c t

Article history: Received 5 June 2015 Received in revised form 11 August 2015 Accepted 13 August 2015 Available online 21 August 2015

The effects of vibration on friction reduction of silicon, aluminum-coated silicon, and aluminum plate sliding against a silicon nitride ball were assessed under various conditions. It was found that friction coefficient could be decreased significantly to 0.1 for all specimens at specific vibration frequency ranges. Also, vibration in the vertical direction resulted in the highest friction reduction of silicon compared to the longitudinal and tangential directions. As for the wear characteristics, it was found that the least amount of wear occurred on silicon followed by aluminum coating and aluminum plate specimens. Though the amount of wear depended on the vibration frequency, the frequency that led to low friction did not necessarily correspond to low wear. & 2015 Elsevier Ltd. All rights reserved.

Keywords: Friction Wear Ceramic Sliding

1. Introduction It has been known since decades ago that imparting vibration to a mechanical component can lead to reduction in friction. In one of the very early works on this topic, Fridman and Levesque used sonic vibration in 1959 to decrease static friction between two metals in contact [1] and Godfrey showed that vibration can effectively reduce friction between two sliding metals in 1967 [2]. In this research, three fixed balls were made to slide against a flat steel plate which was vibrated at relatively low frequencies up to 1000 Hz. Also, Lenkiewicz investigated the effect of vibration on static and kinetic friction coefficients using steel and cast iron [3]. It was shown that friction could be reduced by 80% using relatively low vibration frequency (20–120 Hz) and relatively high vibration amplitude (5–40 μm). Even though the vibration frequency was relatively low, high vibration amplitude allowed for the significant reduction of friction coefficient. This work also suggested that induced vibration can reduce the stick–slip phenomenon. Since the reporting of these early works that demonstrated the effectiveness of vibration to reduce both static and dynamic friction of metals, numerous other works followed [4–8]. Skare and Stahl investigated the static and dynamic frictional behavior of stainless steel specimens under the influence of external vibration using various conditions [4]. They proposed the reason for the decrease in friction under vibrating condition as the separation of the surfaces due to vibration and alteration of the surface properties due to repeated contact. Chowdhury and Helali focused on the

n

Corresponding author. Tel.: þ 82 2 2123 2822; fax: þ 82 2 365 0491. E-mail address: [email protected] (D.-E. Kim).

http://dx.doi.org/10.1016/j.triboint.2015.08.025 0301-679X/& 2015 Elsevier Ltd. All rights reserved.

effects of vibration frequency and amplitude on friction reduction of mild steel sliding against various polymeric materials [5]. The results showed that vibration with frequency and amplitude in the range of  500 Hz and  200 μm, respectively, was effective in friction reduction of polymeric materials. Chowdhury and Helali also investigated the effect of humidity in friction reduction using vibration for mild steel [8]. The results showed the slightly higher reduction in friction could be achieved at high humidity condition (RH  80%) when the vibration frequency was relatively low ( 100 Hz). However, for relatively high vibration frequency ( 500 Hz) humidity did not affect the friction reduction effect. Though vibration for friction reduction is commonly induced in the vertical direction, the effects of vibration in the other directions were also investigated. Gutowski and Leus used an ultrasonic generator to investigate the effects of longitudinal and tangential vibrations in the frequency range of 6–42 kHz on the friction between steel and cast iron [6]. The results showed that tangential vibration, which is perpendicular to the sliding direction, could significantly decrease the driving force of the slider. Teidlet et al. used a specimen integrated with a piezoactuator to investigate the effects of vibration directions (x, y, and z) on the friction of various metals [7]. In this test longitudinal (x) direction showed best friction reduction effect. Thus, the effect of vibration direction on the friction reduction effect varied among these studies depending on the operating conditions such as applied load and vibration amplitude. In addition to the experimental works on the effect of vibration on reduction of friction, numerical modeling of this phenomenon has also been carried out [9–11]. Tworzydol and Becker studied the influence of forced vibration on static friction coefficient by numerical modeling of the compliant contact interface [9]. They showed that frictional force decreased in vibration condition and

S.-S. Yoo, D.-E. Kim / Tribology International 94 (2016) 198–206

explained that the interface damping weakened the friction reduction effects. This study demonstrated the significance of vibration input in dictating the mechanical behavior of contacting asperities, which ultimately affects the frictional behavior. Hess et al. modeled a continuously sliding Hertzian contact system as a non-linear mass–spring–damper system to study the friction reduction effect using vibration [10]. It was suggested that contact loss due to vibrating motion induce reduction of the frictional force. Recently, molecular dynamics (MD) simulation technique was used to investigate the friction reduction effect in vibration condition. Capozza et al. used MD simulation to investigate the role of tiny vibrations in tribological response of a confined system under shear [11]. As these works have shown, the effectiveness of vibration in reducing friction has been demonstrated for sliding systems ranging from macro- to molecular-scale. From the application point of view, friction reduction technique using vibration has been applied in mechanical machining and metal forming processes. In mechanical machining such as cutting and grinding, artificially applied mechanical vibration was used to increase the efficiency of the processes [12–14]. For instance, Kim and Loh showed that 2-dimensional vibration assisted cutting can decrease the cutting force and improve the machining quality [13]. Also, in metal forming process such as the rolling, extrusion and drawing processes, ultrasonic vibration can decrease the frictional force [15–17]. Mousavi et al. investigated the effect of ultrasonic vibration applied on the die during the extrusion process and found that it was effective in reducing the extrusion force [16]. As such, the advantage of using vibration to reduce friction between two materials in contact can be effectively exploited in manufacturing processes where energy savings are highly desirable. As mentioned above, the effectiveness of vibration in reducing friction has been well documented. However, most of the previous works have concentrated on metals and polymers. Only a limited number of studies could be found on the topic of friction reduction of silicon-based materials using vibration. The focus of these works was concerned with increasing the manufacturing process efficiency of these materials [18,19]. Since the frictional behavior under a given vibration condition is known to depend on the stiffness of the contact interface, the effect of vibration on friction is expected to depend on the type of material. In this regard, the motivation of this work was to assess the effectiveness of vibration in reducing the friction of siliconbased materials. Silicon was selected as the target material since it is widely used for fabrication of micro-scale systems [20,21]. Unlike macro-scale systems in which bearings and lubrication provide excellent means to reduce friction, they cannot be employed for

199

micro-scale systems because of limited size and surface tension effect of liquid lubricants [22–25]. Therefore, other techniques have been proposed for micro-scale systems such as thin-film solid lubrication and vapor phase lubrication (VPL) [26–32]. However, it is difficult to deposit solid lubricant coatings on the side walls of micro-scale structures with very small gaps. Also, VPL technique requires continuous supply of vapor phase lubricant which may lead to contamination issues. Hence, other methods, such as vibration, that can lead to reduction in friction of silicon-based materials are needed. In this regard, the aim of this work was to gain fundamental understanding of the frictional behavior of such materials with vibration excitation. Sliding experiments with vibration were performed using a custom-built device. Silicon was used as the test material to slide against a silicon nitride ball. Also, aluminum specimens were used to compare the friction and wear characteristics with those of silicon. Experiments were performed using various vibration frequencies and amplitudes to assess the effects of these variables on friction reduction behavior. Surface damage incurred by the specimens due to vibration was also analyzed. The following sections describe the details of the experimental work.

2. Experimental details 2.1. Specimen preparation In order to assess the effects of vibration on reduction of friction, three types of specimens were prepared: silicon, aluminum-coated silicon, and aluminum plate. Silicon specimens were cut from Si (100) wafers to a dimension of 10  10 mm2. In order to compare the frictional behavior of silicon with a metal under vibration conditions, aluminum was selected. Aluminum specimens were prepared as a coating as well as a bulk material. For the coated aluminum specimen, E-beam evaporator (ULVAC) was used to deposit aluminum on a 10  10 mm2 silicon substrate with a thickness of about 100 nm. The bulk aluminum specimens were cut from a pure aluminum plate to a dimension of 10  10 mm2. For a given specimen, up to 10 tests were conducted with sufficient distance between the sliding tracks. Also, since the specimen surface was quite uniform, there was no reason to expect any difference in the frictional behavior with respect to the location of the ball on the specimen. The average surface roughness of silicon, aluminum-coated silicon, and aluminum plate specimens were  1 nm,  10 nm, and  0.3 μm, respectively. The hardness of silicon was  9.8 GPa and that of the aluminum specimens was about 167 MPa [33,34].

Fig. 1. Experimental set-up for friction measurement with vibration excitation: (a) photograph and (b) schematic image of the set-up.

200

S.-S. Yoo, D.-E. Kim / Tribology International 94 (2016) 198–206

Table 1 Specifications of the piezoelectric actuator used for vibration generation (Noliac plate actuator). Size (width  depth  height) Maximum operating voltage Free stroke

at 200 V at 100 V

Blocking force Capacitance Stiffness Maximum operating temperature Cure temperature Unloaded resonance frequency

10  10  2 mm3 200 V 3.2 mm7 15% 1.6 mm 715% 4000 N7 20% 440 nF7 15% 1250 N/mm7 20% 200 1C 350 1C 4 500 kHz

Silicon nitride ball with a diameter of 2 mm was used as the counter surface. This material was chosen because of its relatively high hardness. The average surface roughness and hardness of the silicon nitride ball was  0.3 μm and  19.6 GPa, respectively [33]. 2.2. Experimental set-up Friction experiments were conducted using a custom built pinon-reciprocating type of a tribotester shown in Fig. 1(a). The silicon and aluminum specimens were used as the plate and the silicon nitride ball was used as the counter surface. Two piezoelectric actuators were used to vibrate the plate specimen in the desired direction. In Fig. 1(b), piezoelectric actuator (I) was used to generate vertical vibration and the other piezoelectric actuator (II) was used to generate longitudinal or tangential vibration. The orientation of the suspension with respect to the actuator (II) was rotated 90 degrees to switch between longitudinal and tangential vibration. The insulation plates made from a silicon wafer were attached to the top and bottom faces of the piezoelectric actuators. This prevented conduction of electricity as well as heat from the actuators to the specimen. A 2 mm diameter silicon nitride ball was attached to a suspension that was connected to a friction force sensor. The sensor was used to record the frictional force using a data acquisition system during the sliding tests. Table 1 shows the specifications of the piezoelectric actuator. 2.3. Resonance frequency and stiffness characteristics of measurement system In order to better understand the frictional behavior of the specimens under vibration conditions, the dynamic characteristics of the measurement system were analyzed. Fig. 2 shows the photograph images of the friction force sensor and the suspension to which the ball was attached. According to the product specification of the friction force sensor, the load range was 750 gf with a resonance frequency of 3.2 kHz. The suspension used to apply the load was taken from a hard disk drive. The vertical stiffness of the suspension was measured to be 20 N/m. The friction force sensor and the suspension were made of aluminum alloy and stainless steel, respectively. Resonance frequency characteristics of the measurement system consisting of the friction force sensor, suspension, and the ball were analyzed using a commercial software. Fig. 3 shows the dynamic characteristics of the system at resonance frequencies and Table 2 lists the values for 1st–6th modes. It could be determined that the 1st, 3rd, and 5th modes were related to the vertical direction vibration of the suspension while the other modes were related to a more complex type of vibration. As for the resonance mode that affected the bending vibration of the friction force sensor, it was the 4th mode with a frequency of 2.7 kHz. Thus, the resonance frequency of the friction force sensor was decreased by about 15% from the original value of

3.2 kHz. The reason for this decrease was due to the increase in the mass of the system by addition of the suspension and the ball. In order to assess the stiffness of the measurement system, the deformation behavior of the system under a static loading condition was analyzed in the vertical direction. Fig. 4 shows the deformation of the measurement system when a vertical load of 1 gf was applied on the bottom center of the ball. It was found that deformation was dominant in the suspension rather than the friction force sensor. The total deformation of the suspension was 501 μm, which corresponded to a vertical stiffness of 20 N/m. Considering that the maximum vibration displacement of the piezoactuator was 1 μm, the vertical vibration motion of the ball was not expected to cause a significant change in the applied load. 2.4. Experimental method Experiments were conducted under various conditions to assess the effects of vibration on the friction reduction of silicon, aluminum-coated silicon, and aluminum plate specimens. For the silicon specimens, vibrations were applied in three directions. Fig. 5 shows the longitudinal, tangential, and vertical vibration directions with respect to the sliding direction. For aluminum-coated silicon and aluminum plate specimens, only the vertical vibration test was conducted because it was identified that this direction had the greatest effect on friction reduction of silicon. Thus, it was logical to compare the frictional behaviors of the aluminum specimens with that of the silicon specimen in the vertical direction. Vibration was applied to the plate specimen by controlling the frequency and amplitude of the sinusoidal input signal. The signal was generated using a function generator (Agilent 33210A 10 MHz function generator) with arbitrary waveform generation capability and a voltage amplifier (FLC P200). The operating frequency of the actuator for the friction tests was set from 0 Hz to 10 kHz. The displacement of the actuator was determined by the voltage amplitude of the input signal. The applied voltage amplitude for the friction tests was set from 0 V to 100 V. At 100 V, the displacement of the piezoelectric actuator was 1.6 mm. Sliding tests were performed in dry conditions with a normal load of 10 mN. The reciprocating stroke and frequency were set to 5 mm and 1 Hz, respectively. The friction coefficient was recorded using a personal computer with respect to the sliding distance under various vibration conditions. All the tests were conducted in ambient environment at room temperature with 30–44% relative humidity. The friction coefficients were measured with respect to the vibration frequency by using the sweep function of the function generator. The operating frequency of the actuator was varied linearly from 0 Hz to 10 kHz over a period of 200 s at a set voltage. The operating voltage was set to 0, 20, 40, 60, 80, and 100 V. Since friction coefficient typically varies with sliding distance at the initial stage of sliding, the friction coefficient measurement was started after the running-in process at which point the friction coefficient value reached a steady state. Essentially, the end of running-in was determined based on the number of cycles that led to a steady state frictional behavior in the preliminary tests. Each test was repeated three times under the same conditions to assure reliability of the data. In order to assess the wear characteristics of the specimens, wear tests were conducted at a fixed operating voltage of 100 V under various operating frequencies. 100 V was selected because it provided the maximum displacement of the actuator. Five different vertical vibration frequency conditions were selected: no vibration, 1 kHz, 5 kHz, 8 kHz, and 10 kHz. For each experiment, the wear test was conducted on a new track on a given specimen. Since wear normally occurs from the onset of sliding, running-in was not performed in the wear tests. Following the wear test, the degree of wear was analyzed using an optical microscope (Bimeince S39A).

S.-S. Yoo, D.-E. Kim / Tribology International 94 (2016) 198–206

201

Fig. 2. Photograph of measurement system components: (a) friction force sensor and (b) suspension.

3. Experimental results 3.1. Frictional behavior of silicon specimen with respect to vibration direction Experiments were performed to investigate the effects of vibration direction on the friction reduction of silicon specimen. Fig. 6 shows the friction coefficient results of vertical, longitudinal, and tangential vibrations with respect to frequency (between 0 and 10 kHz) at operating voltages of 0, 20, 40, 60, 80, and 100 V. Though the data showed some fluctuation, the general trend of frictional behavior between the experimental conditions could be differentiated. Such fluctuation in the friction coefficient behavior was expected for systems with particles and asperities present at the sliding interface between two surfaces [35]. In the case of the longitudinal direction vibration (Fig. 6(a)), operating voltage up to 20 V had no effect on the friction coefficient for the entire frequency range. The friction coefficient was relatively high, ranging between 0.9–1.1. As the voltage increased to 40 V and above, the friction coefficient was reduced significantly depending on the range of vibration frequency. However, the amount of friction reduction did not vary directly with respect to the operating voltage. The lowest friction coefficient of 0.25 was obtained at 60 V and about 7.4 kHz longitudinal vibration frequency. For 100 V, the friction coefficient was somewhat higher, but it was still only about 50% of the 0 or 20 V case. The range of longitudinal vibration frequency over which friction reduction occurred was between 6–8 kHz. In the case of the tangential direction vibration (Fig. 6(b)), the overall friction reduction effect was superior compared with that of the longitudinal direction. The lowest friction coefficient of  0.2 could be obtained with 100 V operating voltage at certain frequencies. Also, relatively low friction coefficients could be achieved over a relatively wide tangential vibration frequency range of 6–10 kHz. As the voltage decreased, the degree of friction reduction also decreased. For operating voltage of 0 and 20 V, there was no friction reduction effect as was the case with the longitudinal vibration. The largest degree of friction reduction was obtained with the vertical direction vibration. For vertical direction vibration (Fig. 6(c)), the friction coefficient decreased drastically to as low as 0.1, at all operating voltages greater than 20 V. Also, the range of vertical vibration frequency over which friction reduction occurred increased with increasing voltage. For 100 V, this range was between 1–9 kHz. Thus, among the three vibration directions, vertical direction resulted in the best overall performance in friction reduction of silicon specimen. It should also be noted that for input voltages of 80 V and 100 V, the friction coefficient started to increase for frequencies above  9 kHz.

3.2. Frictional behaviors of aluminum-coated silicon and aluminum plate specimens Aluminum specimens were used to compare the degree of friction reduction and surface damage due to vibration with that of the silicon specimen. Vertical direction vibration which showed the highest friction reduction effect for silicon was also used for the aluminum specimens. Fig. 7 shows the friction coefficient data of the aluminum plate specimen with vibration in the vertical direction with operating voltages of 0, 20, 40, 60, 80, and 100 V. The friction coefficient decreased drastically to as low as  0.1, at all operating voltages greater than 40 V. Also, the range of vertical vibration frequency over which friction reduction occurred increased with increasing voltage. For 100 V, this range was between 4.5–10 kHz. Fig. 8 shows the friction coefficient data of the aluminum-coated silicon specimen with vibration in the vertical direction with operating voltages of 0, 20, 40, 60, 80, and 100 V. As in the case of the aluminum plate specimen, the friction coefficient decreased drastically to as low as  0.1, at all operating voltages greater than 40 V. Also, the range of vertical vibration frequency over which friction reduction occurred increased with increasing voltage. For 100 V, this range was between 0.5–10 kHz, which was much wider than that of the aluminum plate specimen, and similar to the low friction frequency range of the silicon specimen (1–10 kHz). The difference in the low friction frequency range between the aluminum plate and aluminum-coated specimens may be due to the surface roughness effect. It may be presumed that if the surface is smooth, only a slight separation between the asperities of the two surfaces can lead to reduction in friction. The average surface roughness of the aluminum plate and aluminum-coated silicon specimens was  0.3 μm and  10 nm, respectively. Thus, the smoother aluminum-coated silicon specimen is expected to be more favorable for friction reduction using the vibration method. 3.3. Wear characteristics of the silicon, aluminum-coated silicon, and aluminum plate specimens The wear behaviors of the specimens with respect to vibration conditions were investigated under various vibration frequencies at a set voltage input of 100 V. Fig. 9 shows the optical microscope images of the silicon specimens after the wear tests. As the vibration frequency increased, the surface damage also increased. It was interesting to note that surface damage could not be found after the wear tests with frequencies 0, 1 kHz, and 5 kHz (Fig. 9(a)– (c)). For the 8 kHz wear test, evidence of surface damage could be found (Fig. 9(d)) and for the 10 kHz were test, the surface was

202

S.-S. Yoo, D.-E. Kim / Tribology International 94 (2016) 198–206

Fig. 3. Dynamic characteristics of the measurement system showing the (a) 1st, (b) 2nd, (c) 3rd, (d) 4th, (e) 5th, and (f) 6th resonance mode.

damaged severely (Fig. 9(e)). It should be mentioned that the degree of wear incurred on the specimens was based on the observation made using an optical microscope. Therefore, other high resolution surface characterization methods may be able to detect evidence of wear on the surface of the silicon specimens shown in Fig. 9(a) and (b). Nevertheless, assessment of wear using optical microscopy was sufficient to compare the overall wear

characteristics of the specimens with respect to the vibration conditions. Fig. 10 shows the optical microscope images of the 100-nm-thick aluminum-coated silicon specimens after the wear tests. Unlike the case of the silicon specimen, wear tracks could be distinctly found in all cases. The widths of the wear tracks were quite similar with a dimension of  50 μm. However, the wear particle distribution

S.-S. Yoo, D.-E. Kim / Tribology International 94 (2016) 198–206

Table 2 Resonance frequencies and maximum displacement values of the measurement system. Frequency (Hz)

Maximum displacement (μm)

1 2 3 4 5 6

186 712 1654 2698 3830 4459

280 452 349 76 343 650

0V 20 V 40 V 60 V 80 V 100 V

1.5

Friction coefficient

Mode

203

1.0

0.5

0.0 0

2

4

6

8

10

Longitudinal vibration frequency (kHz)

0V 20 V 40 V 60 V 80 V 100 V

Friction coefficient

1.5

Fig. 4. Deformation of the measurement system for an applied load of 1 gf in the vertical direction.

1.0

0.5

0.0 0

2

4

6

8

10

Tangential vibration frequency (kHz)

0V 20 V 40 V 60 V 80 V 100 V

Fig. 5. Schematic showing the longitudinal, tangential, and vertical vibration directions with respect to the sliding direction.

behavior in the wear track region could be differentiated between the wear tests conducted at 0, 1 kHz, and 5 kHz and those conducted at 8 kHz and 10 kHz. Specifically, for the lower frequency tests, the wear particles were mostly found outside of the wear track whereas for the higher frequency tests, the wear particles were mostly found inside the wear track. This observation was attributed to the contact behavior between the silicon nitride ball and the aluminum-coated silicon specimen with respect to vibration frequency. At low frequencies, as in the case of no vibration, the wear particles were swept to the surrounding area of the wear track by the ball as it slid against the plate specimen. On the other hand, at high frequencies such wear particle sweeping effect was not significant due to the intermittent contact between the ball and the plate specimen. Essentially, the “hoping” motion of the ball compacted the wear particles along the wear track rather than sweeping them out of the track. Fig. 11 shows the optical microscope images of the aluminum plate specimens after the wear tests. In the figure, the vertical lines correspond to the polishing marks. As in the case of the aluminum-coated silicon

Friction coefficient

1.5

1.0

0.5

0.0 0

2

4

6

8

10

Vertical vibration frequency (kHz) Fig. 6. Friction coefficient of the silicon specimen with respect to vibration frequency at operating voltages of 0, 20, 40, 60, 80, and 100 V for (a) longitudinal direction vibration, (b) tangential direction vibration, and (c) vertical direction vibration.

specimen, distinct wear tracks could be found in all cases. The wear track widths for the 8 kHz and 10 kHz tests were larger than those of the other frequencies. Also, two distinctly different wear particle distribution behaviors occurred with respect to the vibration frequency range for the same reason as explained above.

4. Discussion The experimental results showed that vibration was effective in reducing the friction of both silicon and aluminum specimens.

204

S.-S. Yoo, D.-E. Kim / Tribology International 94 (2016) 198–206

0V 20 V 40 V 60 V 80 V 100 V

Friction coefficient

1.5

1.0

0.5

0.0 0

2

4

6

8

10

Vertical vibration frequency (kHz) Fig. 7. Friction coefficient of the aluminum plate specimen with respect to vertical vibration frequency at operating voltages of 0, 20, 40, 60, 80, and 100 V.

Friction coefficient

1.5

0V 20 V 40 V 60 V 80 V 100 V

1.0

0.5

0.0 0

2

4

6

8

10

Vertical vibration frequency (kHz) Fig. 8. Friction coefficient of the aluminum-coated silicon specimen with respect to vertical vibration frequency at operating voltages of 0, 20, 40, 60, 80, and 100 V.

Friction could be reduced from  1.0 to as low as 0.1 at certain voltage and frequency range. The basic mechanism of friction reduction in vibrating systems is known to be the vibration motion of the specimens that cause a momentary separation of the contact during the sliding motion [2]. Thus, friction is expected to be lower if the portion of contact separation event is greater over a given sliding distance. Furthermore, the extent of contact separation would depend on the input voltage and frequency of the actuator. Voltage is related to the displacement amplitude of the actuator and frequency is related to the dynamic characteristics of the system. Measurement of the ball displacement with respect to frequency for various input voltages revealed that it was in the range of several micrometers. From the experimental results, it may be stated that higher voltage allowed for greater probability of contact separation, and thus, low friction could be achieved at relatively high voltage inputs. As for the frequency, it was found that low friction could be achieved for frequencies greater than  5 kHz in the vertical direction. This frequency range could be correlated to the dynamic characteristics of the suspension to which the silicon nitride ball was attached. The modal analysis performed for the measurement system consisting of the friction force sensor, hard disk suspension, and the ball showed that the deformation behavior was dominant in the suspension, which was the most compliant component of the measurement system (Fig. 3). Other works have also shown that hard disk drive suspension vibration response includes bending, torsional, and sway modes [36– 38]. With respect to the vibration direction, vertical bending mode such as the 1st, 3rd, and 5th modes were relevant to the vertical direction vibration. Essentially, the vertical displacement amplitude

Fig. 9. Optical microscope image of the silicon specimen surface after wear test at an operating voltage of 100 V and vertical vibration frequency of (a) 0 kHz, (b) 1 kHz, (c) 5 kHz, (d) 8 kHz, and (e) 10 kHz.

could be maximized by the vertical bending mode. Based on this thought, one may expect that a reduction in friction may occur at an input frequency as low as 188 Hz which corresponds to the 1st bending mode of the measurement system (Table 2). However, experimental results showed that friction reduction generally occurred for frequencies greater than  5 kHz, which was slightly higher than the 6th mode resonance frequency of  4.7 kHz. This discrepancy may be explained from the fact that the dynamic analysis was performed without consideration of the contact stiffness. Since the overall stiffness of the system is expected to increase significantly under contact condition, the fact that the friction reduction occurred at a much higher frequency than the 1st resonance frequency was considered to be feasible. Thus, given the situation that maximum vertical displacement can be achieved at the bending mode natural frequency of the suspension, other factors also affect the actual vertical displacement of the ball sliding against the plate specimen. Particularly, the contact stiffness which is influenced by the material properties is expected to play a significant role as previously reported by Tworzydol and Becker [9]. Also, the state of contact would be affected by the relative roughness of the two contacting surfaces, which in turn would have an influence of the contact stiffness. Considering these factors, it may be postulated that the reason for the different frictional behaviors obtained for the two aluminum specimens was because of the significant difference in the surface roughness as well as the effective stiffness of the material. The aluminum-coated silicon specimen had a much smoother surface and the effective stiffness

S.-S. Yoo, D.-E. Kim / Tribology International 94 (2016) 198–206

Fig. 10. Optical microscope image of aluminum-coated silicon specimen surface after wear test at an operating voltage of 100 V and vertical vibration frequency of (a) 0 kHz, (b) 1 kHz, (c) 5 kHz, (d) 8 kHz, and (e) 10 kHz.

was expected to be higher than that of the aluminum plate specimen because of the hard and stiff silicon substrate below the coating. Thus, even for the same material, friction reduction behavior with respect to vibration differed significantly. As for the effect of vibration direction on the extent of friction reduction, it was found that vertical direction exhibited the lowest friction compared with the other two directions. This was because vertical displacement was more prone to cause intermittent contact separation between the silicon nitride ball and the silicon specimen. However, it should be noted that in the case of metal forming process vibration in the longitudinal and tangential directions are utilized to lower friction rather than vertical vibration [16,17]. This is because in metal forming there is usually no gap between the die and the workpiece in the vertical direction, and thus, separation of the asperity contact by vertical vibration is not feasible. Essentially, due to the high vertical stiffness of the metal forming process, vibration only in the longitudinal and tangential directions can be applied. In addition to the frictional behavior of the specimens, wear characteristics were also analyzed. Overall, the aluminum specimens showed significantly higher wear compared to the silicon specimen for vibration frequencies below 10 kHz. Though some degree of wear was observed with the silicon specimen tested at a relatively high frequency of 8 kHz, the degree of surface damage was much less than that of the aluminum specimens. The reason for the relatively low wear incurred by the silicon specimen was attributed to its high hardness compared to the aluminum specimens. Also, the effect of

205

Fig. 11. Optical microscope image of aluminum plate specimen surface after wear test at an operating voltage of 100 V and vertical vibration frequency of (a) 0 kHz, (b) 1 kHz, (c) 5 kHz, (d) 8 kHz, and (e) 10 kHz.

vibration frequency on wear was quite significant in the case of the silicon specimen. Particularly, as the frequency increased from 8 kHz to 10 kHz, the amount of wear increased drastically. This indicated that the surface durability threshold of the silicon specimen was exceeded at a frequency between these two values. As for the relationship between friction and wear, low friction did not necessarily result in low wear and high friction did not always result in high wear. For example, even though significant amount of wear occurred in all the specimens at 100 V and 10 kHz vibration frequency, friction was still quite low. On the other hand, in the case of the silicon specimen 1 kHz vibration frequency resulted in almost no wear but high friction. The reason for this outcome is not clear but it may be assumed to be due to the contrasting effects of the “hopping” motion on friction and wear. Essentially, it may be presumed that “hopping” motion of the ball was beneficial for reducing friction by contact separation but adverse from the view point of wear since repeated impact due to the “hopping” motion can increase the instantaneous contact stress. Thus, even though wear may be high due to greater surface impact of the ball, the larger contact separation aided in reduction of friction. Overall, the best friction and wear behavior was obtained with the silicon specimen with vertical vibration at 100 V and 5 kHz frequency. For this case, the friction coefficient was  0.1 and wear was not detected. This outcome was significantly better than the aluminum specimens, which indicated that vibration method can be effectively utilized to reduce friction and wear of silicon for potential applications in micro-systems.

206

S.-S. Yoo, D.-E. Kim / Tribology International 94 (2016) 198–206

5. Conclusion The effects of vibration frequency and amplitude on friction reduction and wear characteristics of silicon, aluminum-coated silicon, and aluminum plate sliding against a silicon nitride ball were assessed. Also, the effect of vibration direction on friction reduction was investigated using the silicon specimen. Based on the experimental results, the following conclusions may be drawn: 1) Vibration method was effective in significantly reducing the friction coefficient of all specimens at certain frequency ranges. Generally, higher actuator input voltage resulted in lower friction over a wider frequency range. 2) In the case of the silicon specimen, vertical direction vibration showed the greatest reduction in friction compared with the longitudinal and tangential directions. The friction coefficient could be reduced from  1.0 for no vibration to as low as  0.1 at 100 V over a relatively wide range of vertical vibration frequency. For the other two directions, friction coefficients of 0.2–0.25 could be obtained. 3) In the case of the aluminum specimens the friction coefficient could be reduced from  0.8 for no vibration to as low as  0.1 at 100 V. The aluminum-coated silicon specimen showed a wider frequency range over which friction reduction effect occurred compared to the aluminum plate specimen. 4) Wear of the silicon specimen was significantly lower than the aluminum specimens for vibration frequencies below 10 kHz. Particularly for frequencies below 8 kHz, wear could not be detected on the silicon specimen. However, at 10 kHz, evidence of severe wear was found, which suggested that surface durability threshold frequency for silicon lies between 8 and 10 kHz. 5) Overall, vertical vibration at 100 V and 5 kHz frequency using the silicon specimen resulted in the best friction and wear behavior. A low friction coefficient of  0.1 could be achieved with no detectable wear. This outcome indicated that vibration method can be effectively utilized to reduce friction and wear of silicon.

Acknowledgment This work was supported by the National Research Foundation of Korea (NRF) Grant funded by the Korea Government (MSIP) (No. 2010-0018289). References [1] Fridman HD, Levesque P. Reduction of static friction by sonic vibrations. J Appl Phys 1959;30:1572. [2] Godfrey D. Vibration reduces metal to metal contact and causes an apparent reduction in friction. Tribol Trans 1967;10:183–92. [3] Lenkiewicz W. The sliding friction process: effect of external vibrations. Wear 1969;13:99–108. [4] Skare T, Stahl JE. Static and dynamic friction processes under the influence of external vibrations. Wear 1992;154:177–92. [5] Chowdhury MA, Helali MM. The effect of amplitude of vibration on coefficient of friction for different materials. Tribol Int 2008;41:307–14. [6] Gutowski P, Leus M. The effect of longitudinal tangential vibrations on friction and driving force in sliding motion. Tribol Int 2012;55:108–18. [7] Teidelt E, Starcevic J, Popov VL. Influence of ultrasonic oscillation on static and sliding friction. Tribol Lett 2012;48:51–62.

[8] Chowdhury MA, Helali MM. The effect of frequency of vibration and humidity on coefficient of friction. Tribol Int 2006;39:958–62. [9] Tworzydlo WW, Becker E. Influence of forced vibrations on the static coefficient of friction – numerical modeling. Wear 1991;143:175–96. [10] Hess DP, Soom A, Kim CH. Normal vibrations and friction at a hertzian contact under random excitation: theory and experiments. J Sound Vib 1992;153:491–508. [11] Capozza R, Vanossi A, Vezzani A, Zapperi S. Triggering frictional slip by mechanical vibrations. Tribol Lett 2012;48:95–102. [12] Ko JH, Tan SW. Chatter marks reduction in meso-scale milling through unltrasonic vibration assistance parallel to tooling's axis. Int J Precis Eng Manuf 2013;14:17–22. [13] Kim GD, Loh BG. Cutting force variation with respect to tilt angle of trajectory in elliptical vibration V-grooving. Int J Precis Eng Manuf 2013;14:1861–4. [14] Zhong ZW, Rui ZY. Grinding of single-crystal silicon using a micro-vibration device. Mater Manuf Process 2005;20:687–96. [15] Pasierb A, Wojnar A. An experimental investigation of deep drawing and drawing processes of thin – walled products with utilization of ultrasonic vibrations. J Mater Process Technol 1992;34:489–94. [16] SAAA Mousavi, Feizi H, Madoliat R. Investigations on the effects of ultrasonic vibrations in the extrusion process. J Mater Process Technol 2007;187188:657–61. [17] Siegert K, Ulmer J. Superimposing ultrasonic waves on the dies in tube and wire drawing. J Eng Mater Technol 2001;123:517–23. [18] Saffar S, Abdullah A, Gouttebroze S, Zhang ZL. Ultrasound-assisted handling force reduction during the solar silicon wafers production. Ultrasonics 2014;54:1057–64. [19] Ding K, Fu YC, Su HH, Gong XB, Wu KQ. Wear of diamond grinding wheel in ultrasonic vibration-assisted grinding of silicon carbide. Int J Adv Manuf Technol 2014;71:1929–38. [20] Williams JA, Le HR. Tribology and MEMS. J Phys D: Appl Phys 2006;39:201–14. [21] Savkar AA, Murphy KD, Leseman ZC, Mackin TJ. On the use of structural vibration to release stiction failed MEMS. J Microelectromech Syst 2007;16:163–73. [22] WSN Trimmer. Microrobots and micro-mechanical systems. Sens Actuator 1989;19:267–87. [23] Tian X, Bhushan B. The micro-meniscus effect of a thin liquid film on the static friction of rough surface contact. J Phys D: Appl Phys 1996;29:163–78. [24] Kim HJ, Kim DE. Nano-scale friction: a review. Int J Precis Eng Manuf 2009;10:141–51. [25] Kim HJ, Yoo SS, Kim DE. Nano-scale wear: a review. Int J Precis Eng Manuf 2012;13:1709–18. [26] Sung IH, Kim DE. Micro/nano-tribological characteristics of self-assembled monolayer and its application in nano-structure fabrication. Wear 2003;255:808–18. [27] Penkov OV, Lee DH, Kim HJ, Kim DE. Frictional behavior of atmospheric plasma jet deposited carbon-ZnO composite coatings. Compos Sci Technol 2013;77:60–6. [28] Ryu BH, Barthel AJ, Kim HJ, Lee HD, Penkov OV, Kim SH, et al. Tribological properties of carbon nanotube–polyethylene oxide composite coatings. Compos Sci Technol 2014;101:102–9. [29] Penkov OV, Kim HJ, Kim HJ, Kim DE. Tribology of graphene: a review. Int J Precis Eng Manuf 2014;15:577–85. [30] Asay DB, Dugger MT, Kim SH. In-situ vapor phase lubrication of MEMS. Tribol Lett 2008;29:67–74. [31] Kim SH, Asay DB, Dugger MT. Nanotribology and MEMS. Nano Today 2007;2:22–9. [32] Yoo SS, Kim DE. Vapor phase lubrication using high molecular weight lubricant for friction reduction of metals. Int J Precis Eng Manuf 2014; 15:867–73. [33] Bhushan B, Li X. Micromechanical and tribological characterization of doped single-crystal silicon and polysilicon films for microelectromechanical devices. J Mater Res 1997;12:54–63. [34] Totten GE, MacKenzie S. Handbook of aluminum. Physical metallurgy and processes, vol. 1. Boca Raton: CRC Press; 2003. [35] Kim DE, Cha KH, Sung IH, Bryan J. Design of surface micro-structures for friction control in micro-systems applications. CIRP Ann – Manuf Technol 2002;51:495–8. [36] Pang CK, Lewis FL, Ge SS, Guo G, Chen BM, Lee TH. Singular perturbation control for vibration rejection in HDDs using the PZT active suspension as fast subsystem observer. IEEE Trans Ind Electron 2007;54:1375–86. [37] Jun S, Zhaowei Z. Finite element modeling and modal testing of vibration characteristics of disk drive with a PZT microactuator. Sens Actuators A: Phys 2002;100:257–63. [38] Xu LM, Guo N. Modal testing and finite element modelling of subsystem in hard disk drive. Mech Syst Signal Process 2003;17:747–64.