Volume increase phenomena in reciprocal scratching of polycarbonate studied by atomic force microscopy Andrew Khurshudov and Koji Kato Laboratory of Tribology, Faculty of Engineering, Tohoku University, Sendai 980-77, Japan

~Received 2 February 1995; accepted 1 July 1995! An atomic force microscope ~AFM! was used to study the microwear process of polycarbonate ~PC!. Testing included line scratching of the polymer surface using a microfabricated Si3N4 AFM tip of 10–20 nm radius. Interfacial adhesion, friction, the effect of the number of scratches, and scanning speed were studied. Unlike previous reports, projections as a result of polycarbonate apparent volume increase have been observed after reciprocal scratching on the same track ~without an AFM tip lateral feed!. In spite of analytically predicted elastic/plastic contact, no plastic deformation was found during the tests. A model, based on the formation of cracks and their growth was suggested to explain these phenomena. No viscoelastic/plastic behavior ~scanning speed effect on friction or microwear! was found. © 1995 American Vacuum Society.

I. INTRODUCTION

B. The AFM and AFM tip

An atomic force microscope ~AFM! operating in the contact mode became a popular instrument in the field of microtribology. Unlike macrotribology, microtribology deals with the properties of the fine surface layers of nanometer order thickness. An AFM was used for the study of atomic scale adhesion, friction, and hardness.1–5 It was also used for a wear resistance characterization of the surface layers of hard coatings, single crystals, etc.6,7 Polymeric materials are widely used in tribology because of, among other reasons, their low specific gravity, chemical inertness, and low friction coefficient. The revealing of their wear micromechanisms is also a way to understand their macrotribological behavior. Study of microwear mechanisms of some polymers ~PC, PMMA, epoxy! was previously done by Hamada and Kaneko.8 –10 New phenomenon such as projections formation on the polymer wear surface with a volume increase was found, unlike coatings and other materials which have shown only surface depression. The purpose of this article is to present some new experimental data on the microwear of polycarbonate and to analyze the alternative wear mechanism.

The atomic force microscope SPA300 series ~Seiko Instruments, Inc., Japan! operating in the contact mode was used in this study ~Fig. 1!. This AFM uses a quartered photodetector to pick up a reflected laser beam for measurement of the triangle gold-coated cantilever deformation under loading. Values of the normal and lateral forces were estimated by the AFM software using the data about cantilever bending and torsion spring constants.12 The data for microfabricated cantilever and tip, as provided by the AFM supplier, are presented in Table II. The tip shape was pyramidlike and was sharpened by the etching process,13 which provided a tip radius of 10–20 nm.

II. EXPERIMENT A. Materials

Polycarbonate ~PC!, best known as an engineering plastic, is characterized by the following properties: low density, high strength, stiffness, hardness, and toughness over temperatures ranging from 2150 to 1135 °C. The main disadvantages of PC are: limited chemical resistance, notch sensitivity, and susceptibility to stress cracking.11 PC possesses a low tendency for crystalline formation and is mostly amorphous. This is the reason for its high toughness. All commercial polycarbonates, with the exception of some used for special purposes, have the same chemical composition. Injection molded PC was used in this study. Some selected properties of polycarbonate are presented in Table I. The AFM tip material was Si3N4 in the shape of a pyramid. 1938

J. Vac. Sci. Technol. B 13(5), Sep/Oct 1995

C. Experimental method

The experimental method included reciprocal line scratching of the polycarbonate surface by the AFM tip. The stroke was equal to 1 mm. Scanning frequency, 4 Hz ~;20 mm/s!, nominal normal force ~without taking into account an adhesive force! was 44.5 nN. The scratching direction was parallel to the cantilever axis in order to increase the cantilever’s stiffness. Measurement included topography and friction force distribution on a 350–500 nm2 area around the wear mark. Scanning direction was perpendicular to the cantilever axis to improve its sensitivity to the friction force, which acted in the same direction. During measurements, resolution was

TABLE I. Selected properties of polycarbonate.a Material Specific gravity ~N/m3! Melting temperature ~°C! Glass-transition temperature ~°C! Elastic modulus ~GPa! Poisson ratio Tensile strength ~MPa! Elongation ~%! Yield stress ~MPa! a

Polycarbonate ~PC! 1220 ••• 149 2.1–2.3 ;0.5 55–70 98 .55

References 9 and 11.

0734-211X/95/13(5)/1938/7/$6.00

©1995 American Vacuum Society

1938

1939

A. Khurshudov and K. Kato: Volume increase phenomena in reciprocal scratching

1939

TABLE II. The data for microfabricated cantilever and tip. Cantilever shape Cantilever length ~mm! Cantilever thickness ~nm! Cantilever spring constant ~N/m! Resonance frequency ~kHz! Tip shape Tip material Tip radius ~nm!

Triangle 100 400 0.09 40 Sharpened pyramid Si3N4 10–20

usually 256 lines 3 256 pixels each, scanning frequency was 2 Hz ~;10 mm/s!, normal force was 0.087 nN. Experiments were carried out under the thermo-vibroprotecting cover of the AFM. Polymeric specimen surface was cleaned by 99.5% ethanol just before the test, and was kept under the cover for 5 min to dry. D. Contact conditions FIG. 1. Schematic diagram of the AFM.

Tests included approach ~force! curve measurement to evaluate possible adhesive force in the interface of Si3N4 tip/polycarbonate. Pull-off force was measured many times

FIG. 2. Topography of the surface after ~a! 5, ~b! 20, ~c! 100, and ~d! 700 reciprocal scratches, repeated on the same place. JVST B - Microelectronics and Nanometer Structures

1940

A. Khurshudov and K. Kato: Volume increase phenomena in reciprocal scratching

1940

FIG. 4. Projection dimensions dependence on the number of scratches.

Hertzian contact pressure at which plastic flow on the polymer surface will start:16 Q maxc 51.852Y 51.8523555102 MPa,

~1!

where Y is the tensile yield strength of the polymer ~see Table I!. Maximum Hertzian contact pressure for the sphere/plane contact is16 Q max5@6PE2/p3R2#0.33, where E is the combined elastic modulus and P is the apparent contact force-result of the cantilever elastic bending. For R510 nm, P544.5 nN, E52.9 GPa, we have Q max5897.6 MPa, for R520 nm, Q max5570 MPa. Even without counting the additional

FIG. 3. ~a! Topography and ~b! friction force distribution on the surface after ten reciprocal scratches, repeated on the same place. Positive values correspond to larger friction force.

and was equal to ;20 nN with a relatively small scatter of 3%–5%. According to the Derjaguin–Muller–Toporov ~DMT! theory of elastic contact with adhesion,14,15 maximum pull-off force is 2 p wR, adhesive force P adh in the interface is also equal to 2 p wR. w5 g 1 1 g 2 2 g 12 , g 1 , g 2 are the surface energies of contacting bodies, g 12 is the energy of the new surface formation, and R is the tip radius. Accordingly, we may assume a presence of additional load in the interface of the order of 20 nN even for almost zero load during the measuring of topography ~0.087 nN!. To evaluate contact conditions between the tip and polymer surface we calculated the critical value of maximum J. Vac. Sci. Technol. B, Vol. 13, No. 5, Sep/Oct 1995

FIG. 5. Model of projection formation resulting in apparent volume increase. ~a! Elastic asymmetrical loading and unloading with no damage to the polymer, ~b! first formation of cracks under loading and first extra-volume formation during unloading, ~c! cracks propagation and accumulation under loading and projection growth during unloading.

1941

A. Khurshudov and K. Kato: Volume increase phenomena in reciprocal scratching

FIG. 6. Possible structures of projection: ~a! with many cracks, ~b! with one crack, ~c! with a combination of cracks of different sizes.

adhesive force effect these values are much larger than Q maxc . This corresponds to elastoplastic character of tip/ polymer surface interaction.

Unlike the conclusions of Hamada and Kaneko, we observed volume increase phenomena without lateral feed during reciprocal scratching on the same place. An analysis of test results with the number of cycles of 100 or more allowed for the location of the AFM tip position relative to the projections. Images of the AFM tips are shown in Figs. 2~a!–2~d!. The important fact is that there are no plastic grooves reliably observed on the surface up to at least, 20 cycles, in spite of analytically predicted plastic contact @Figs. 2~a!, 2~b!, and 3~a!#. A possible explanation is: the tip radius is larger than was predicted by the supplier’s data and the tip surface is more flat, resulting in a low contact pressure. The contact area size, necessary to provide elastic contact conditions, may be roughly evaluated assuming the uniform contact stress distribution between tip and surface using the following formula: P total/ p r 2c ,Q maxc ,

III. RESULTS AND DISCUSSION A. Apparent volume increase (projections formation) mechanism

Reciprocal line scratches were made on the PC surface with the different number of cycles. Figures 2~a!–2~d! and 3~a! present topography of the surface after 5, 10, 20, 100, and 700 reciprocal scratches were repeated on the same place. Volume increase on the scratched surfaces can be seen. Surface distortions after less than five scratches were too small to be observed. Figure 4 shows the effect of the number of scratches on both the height and width of the deformed area. Quite detailed study of the PC microwear mechanism was undertaken in previous papers.8 –10 Experimental evidence of the plastic deformation of the polymeric surface which occurred during the line scratch ~formation of groove and ridges! were presented in these papers, but volume increase was observed during scan scratching ~line scratching with the lateral feed! even under nominally elastic contact conditions.9 ‘‘Upheaved’’ areas and projections were found to be softer than the surface layer of the polycarbonate. It was suggested that the volume increase during projection formation may be the result of two possible processes: either voids or crevices formation in the surface layer of the polymer, or polycarbonate mixing with water molecules.8,9 Later, Kaneko and Hamada,10 on the basis of scan scratches with different feeds, concluded that upheaval formation with volume increase is the result of neighbor line-scratches interaction during scan scratch and that a line scratch itself can produce plastic grove and ridges, but not result in volume increase. So it was supposed that the observed volume increase process was based on plastic deformation of the surface and grooves-ridges formation and ridges interaction during the scan scratch. This model cannot explain a volume increase observed under nominally elastic contact conditions.9 In this article we will not distinguish between ‘‘upheaval’’ and ‘‘projection,’’ but rather use the term ‘‘projection’’ to define any volume increase phenomena. JVST B - Microelectronics and Nanometer Structures

1941

~2!

where P total5 P1 P adh>65 nN, r c is the contact spot radius. According to this formula, r c should be of the order of 15 nm and wear groove width should be about 30 nm. Grooves from Figs. 2~c! and 2~d! have an average width of about 20–30610 nm. It corresponds to our evaluations @Eq. ~2!# and allows us to conclude that the contact conditions were elastic rather than plastic. It also means that the surface phenomenon, like the formation of projections, was the result of elastic contact with friction. Height ~and volume! of the deformed surface increased with the maximum ;10 cycles ~Fig. 4!. Both height and width showed fast increase during the first 5–20 cycles with the transition to more slow increase. Beginning at 100 cycles, the groove may be found @Fig. 2~c!#. At 700 cycles, two projections, divided by the groove, can be seen on the surface @Fig. 2~d!#. The groove was not uniform, but periodically disappeared and appeared again every ;150–200 nm. It can be seen from Fig. 2~c! that an interaction of the tip and surface results in asymmetrical surface distortion. Even in the case of two projections, formed on the surface after 700 cycles @Fig. 2~d!#, the left projection is higher than the right. The only reasonable explanation is that the AFM tip was not perpendicular to the surface during scratching, but contacted from some angle. This angle may be a result of initial nonperpendicularity, caused by the tip holder. Our scratch tests were carried out under relatively low sliding speeds of the order of a few mm/s. This means low contact temperatures. Projections appeared on the surface contacted by the AFM tip on a small area of its surface. Under these conditions we may rule out possible tribochemical phenomena, resulting in volume increase, or some mechanically activated processes, like mixing of water molecules with a polymer,8,9 also resulting in a volume increase. Most likely, the only reasonable explanation of the observed phenomena is the mechanical changes in the polymer. Figure 5 presents a possible model of projection formation resulting in apparent volume increase. There are three main steps, according to this model:

1942

A. Khurshudov and K. Kato: Volume increase phenomena in reciprocal scratching

FIG. 7. Profiles of the area from Fig. 3~a! in both ~a! x-axis and ~b!,~c! y-axis directions. J. Vac. Sci. Technol. B, Vol. 13, No. 5, Sep/Oct 1995

1942

1943

A. Khurshudov and K. Kato: Volume increase phenomena in reciprocal scratching

~1! First, asymmetrical elastic loading of the initial surface @Fig. 5~a!# results in asymmetrical distribution of stresses and pileup to one side of the contact which disappears after unloading ~the AFM tip moves out of contact!. The higher the friction, the bigger the pileup. ~2! After some specific number of cycles, shear stresses during loading result in the formation of cracks under the surface of the polymer @Fig. 5~b!#. Elastic recovering after unloading results in cracks surfaces separation, producing additional volume and resulting in local volume increase of the polymer with the formation of projections. ~3! Every additional loading cycle leads to more accumulation, propagation, and interaction of the cracks @Fig. 5~c!#. Elastic recovering during unloading results in further growth of the projection. The formation of relatively deep grooves on the surface after a sufficient number of cycles provides more conformal contact between the AFM tip and the surface. Polymer volumes on the other side of the tip became involved in the reciprocal deformation with the following defects accumulation and growth. In this case, formation of the second smaller projection on the opposite side of the AFM tip is possible @see Fig. 2~d!#. Figures 6~a!– 6~c! show some possible types of projection structure: with many cracks ~a!, one crack ~b!, and a combination of cracks of different sizes ~c!. Taking into account the size of projections, it is almost impossible to detect the real structure directly. Figures 7~a!–7~c! show profiles of the area from Fig. 3~a! in both x-axis and y-axis directions. The surface profile along the scratched area is much more diverse compared to that for the unscratched area, and has some random character. It is similar to what we can expect from the suggested model, based on the formation of cracks, which produces the area full of interacting cracks and voids of different sizes and orientations ~Figs. 5 and 6!. Presented experimental data correspond to the case of the elastic interaction with friction of some rigid tip or abrasive particle of a large radius with the polymer surface. In general, abrasive particles and surface asperities possess nonsymmetrical shapes and their interaction with the surface of the counterbody results in an asymmetrical stress field. Asymmetrical interaction with the polycarbonate surface under elastic contact conditions with friction, as it was shown above, may cause the formation of cracks under the surface ~fatigue! and the formation of projections with apparent volume increase. B. Scanning speed effect

In this test some line scratches were made with different scanning speeds: 0.5, 1, 5, 10, and 20 mm/s. No effect of speed on wear track sizes or shape were observed. C. Friction force

Figure 3 shows the topography of the area scratched with ten cycles @Fig. 3~a!# and the friction force distribution @Fig. 3~b!# made with the normal load equal to 0.087 nN over the same surface. In the case of friction force, positive values JVST B - Microelectronics and Nanometer Structures

1943

correspond to larger friction force in Fig. 3~b!. Comparison of the normal force value with the friction force value ~about 30 nN! gives us an incredibly high friction coefficient of about 337. It is well known that for the relatively small normal loads, the friction coefficient becomes dependent on a normal load and increases with the load decrease. This is related to the effect of adhesive forces between the surfaces, which increases the true contact load.17 If, for example, one calculates the friction coefficient while taking into account the adhesive force of 20 nN as an additional normal load, the friction coefficient becomes equal to 1.5 only. Anyway, high friction force acts at the interface even for a very small normal load, causing high shear stress. To study the effect of scanning speed on friction force value, line-scratch tests with the friction force continuously recording, were done for the speed region of 0.12– 64 mm/s for the fixed normal load. These tests have shown no effect of scanning speed on the friction force value. For example, for the normal load of 10 nN, friction force was constant and equal to 55 nN with a scatter of about 5%–10% along the scanning distance. It shows, that the interfacial processes, responsible for the friction, were independent of speed, like, for example, elastic deformation. Concerning the fact that both friction and surface shape changing were independent of scanning speed in our tests, it should be mentioned, that the same fact was reported before during friction tests using the AFM.18,19 This may be a result of the relatively slow scanning speeds ~up to a few tens of mm/s!, normally used in the AFM. IV. CONCLUSIONS The results of this work may be summarized as follows: ~1! Apparent volume increase ~projection! in the region of reciprocal line scratching on the same place was first found for polycarbonate. ~2! A model, based on the process of the formation of cracks and their growth, was introduced to explain projections formation phenomena under elastic contact conditions. ~3! Friction force and the process of the polymeric surface shape changing were found to be independent of the scratching ~sliding! velocity. 1

C. M. Mate, G. M. McClelland, R. Erlandsson, and S. Chiang, Phys. Rev. Lett. 59, 1942 ~1987!. 2 C. M. Mate, G. S. Blackman, A. B. Jaffe, M. R. Lorenz, V. J. Novotny, and L. L. Sanders, Proceedings of the Japan International Tribol. Conference, Nagoya, Japan, Oct. 29–Nov. 1, 1990 ~unpublished!, p. 1293. 3 T. Miyamoto, R. Kaneko, and Y. Ando, Trans. ASME J. Tribology 112, 567 ~1990!. 4 N. A. Burnham, D. D. Domingues, R. L. Mowery, and R. J. Colton, Phys. Rev. Lett. 64, 1931 ~1990!. 5 C. J. Lu, D. Bogy, and R. Kaneko, Trans. ASME J. Tribology 116, 175 ~1994!. 6 T. Miyamoto, R. Kaneko, and S. Miyake, J. Vac. Sci. Technol. B 9, 1338 ~1991!. 7 T. Miyamoto, S. Miyake, and R. Kaneko, Wear 162–164, 733 ~1993!. 8 E. Hamada and R. Kaneko, J. Phys. D 53–56, 31 ~1992!. 9 E. Hamada and R. Kaneko, Ultramicroscopy 42–44, 184 ~1992!. 10 R. Kaneko and E. Hamada, Wear 162–164, 370 ~1993!. 11 H. Domininghaus, Plastics for Engineers. Materials, Properties, and Applications ~Carl Hanser Verlag, Munich, 1993!, p. 423. 12 R. Overney and E. Meyer, Mater. Res. Soc. Bull. XVIII, 5, 26 ~1993!. 13 E. Meyer and H. Heinzelmann, in Scanning Tunneling Microscopy II,

1944

A. Khurshudov and K. Kato: Volume increase phenomena in reciprocal scratching

Springer Ser. Surf. Sci. Vol. 28 ~Springer, Berlin, 1992!, pp. 99–149. B. V. Derjaguin, V. M. Muller, and Y. P. Toporov, J. Colloid. Interface Sci. 53, 314 ~1975!. 15 J. Maugis, Colloid Interface Sci. 150, 243 ~1992!. 16 B. Bhushan, Tribology and Mechanics of Magnetic Storage Devices ~Springer, Berlin, 1990!, p. 170. 14

J. Vac. Sci. Technol. B, Vol. 13, No. 5, Sep/Oct 1995

1944

I. Etsion and M. Amit, Trans. ASME J. Tribology 115, 406 ~1993!. S. R. Cohen, G. Neubauer, and G. M. McClelland, J. Vac. Sci. Technol. A 8, 3449 ~1990!. 19 G. M. McClelland and J. N. Glosli, in Fundamental of Friction: Macroscopic and Microscopic Processes, edited by I. L. Singer and H. M. Pollock ~Kluwer Academic, Dordrecht, 1992!, pp. 405– 425. 17 18