Journal of Biomechanical Science and Engineering

Transcription

1 Science and Engineering Development of Biomimetic Bearing with Hydrated Materials * Yoshitaka NAKANISHI **, Tatsuki TAKASHIMA ***, Hidehiko HIGAKI ****, Ken SHIMOTO ****, Takatoshi UMENO *****, Hiromasa MIURA and Yukihide IWAMOTO **Graduate School of Science and Technology, Kumamoto University, Kurokami, Kumamoto , Japan y-naka@mech.kumamoto-u.ac.jp ***MECC Co.,Ltd Fukudou, Ogouri, Fukuoka , Japan ****Fuculty of Engineering, Kyushu Sangyo University, Matsukadai, Higashiku, Fukuoka , Japan *****Faculty of Education, Fukuoka University of Education, 1-1 Bunkyoumachi Akama, Munakata, Fukuoka , Japan Graduate School of Medical Sciences, Kyushu University, Maidashi, Higashiku, Fukuoka , Japan Abstract In order to produce a biomimetic bearing made from a hydrated material, a polyvinyl formal (PVF) bearing that mimics articular cartilage has been proposed. PVF has material and tribological properties that reflect the different internal structures resulting from different parameters used in the PVF formation process. Reciprocating and multidirectional wear tests were performed, and plastic deformation attributed to pore deformation was observed; however, microscopic observations did not reveal any damage due to friction. PVF exhibited high wearresistance properties, and it is concluded that the proposed bearing overcomes the problem of low wear-resistance in bearing materials capable of developing hydration lubrication. Key words: Biomimetics, Bearing, Polyvinyl Formal (PVF), Polyvinyl Alcohol (PVA), Articular Cartilage, Wear, Lubrication 1. Introduction *Received 5 Nov., 2008 (No ) [DOI: /jbse.4.249] The highly effective lubrication systems used in animals and humans have been investigated and several attempts have been made to reproduce them artificially [1-3]. The hydration lubrication employed in articular cartilage has frequently been reproduced using gels or porous materials as bearing materials. One such material is polyvinyl alcohol (PVA) hydrogel, which is one of the few polymers with hydrophilic properties. It is anticipated to realize a wide range of clinical applications due to its high water-holding capacity and high biocompatibility [4]. However, a major problem with PVA hydrogel is its low wear resistance. Therefore, in this study, we attempt to improve its wear resistance by performing a chemical cross-linking reaction in PVA. The resulting cross-linked PVA is referred to as polyvinyl formal (PVF) [5]. The possibility of using it as a bearing material is discussed. 249

2 2. Experimental Details 2.1. Polyvinyl formal (PVF) as a Bearing Material The chemical reaction for the formation of PVF from PVA is shown in Fig. 1. When PVA is formalized by formaldehyde (HCHO) in the presence of an acid catalyst (H 2 SO 4 ), it transforms into PVF, becoming water-resistant and acquiring very good mechanical properties. In the formalization reaction, intermolecular formals react with PVA molecules, and intramolecular reactions occur simultaneously, in which formals react with 1-2 glycols and 1-3 glycols within a single molecule. The resulting PVF is a copolymer consisting of a mixture of alcoholic hydroxyl groups, acetyl groups, and formal groups. Fig. 1. Chemical reactions for converting polyvinyl alcohol (PVA) into polyvinyl formal (PVF) for use as a biomimetic bearing material. Fig. 2. Formation of PVF as a continuous porous material. Figure 2 depicts the process of PVF formation on a microscopic scale. The PVA dissolved in water is subjected to intermolecular formalization, which enhances the crosslinks between the principal chains of the polymer. Hydrophilicity is preserved due to the presence of residual alcoholic hydroxyl groups that do not participate in the reaction. Water is removed during the reaction and the intermolecular forces cause the principal chains to aggregate, resulting in the formation of pores and beams. The PVF formed by this process is hydrophilic and has a continuous porous structure. This structure is considered to promote hydration lubrication considerably. The material and tribological properties of PVF are determined by the relationship between the pores and beams shown in Fig. 2. Consequently, if these structural parameters can be controlled in the formation process, it will be possible to produce a suitable material for biomimetic bearings in industrial applications. Figure 3 shows the method used to produce PVF samples. Aqueous solutions of PVA (concentrations: 10 wt% to 15 wt%; mean degree of polymerization: , saponification degree: mol%) was prepared, and allowed to stand for 24 hours to remove air bubbles. Formaldehyde (HCHO) with an acid catalyst (H 2 SO 4 ) for formalization was added to the PVF solution, and this solution was again allowed to stand 250

3 Fig. 3. Method used to produce PVF. for 24 hours to remove air bubbles. The formalization reaction (i.e., cross-linking reaction) of the PVA was accelerated by heating it at 60 ºC for 72 hours. It was then neutralized to allow it to be used in practical applications. The effects of the concentrations of PVA (10-15 wt% in the PVA solution), formaldehyde ( wt%), and sulfuric acid ( wt% to PVA) on the continuous porous structure of the PVF was investigated through cross-sectional SEM observations. The porosity [P w ] and beam density [D d, D w ] of the PVF samples in dry or hydrated state were estimated by gravimetric and volumetric analyses by using these equations [1-3]. These estimations were performed on the assumption that the hydration did not affect the average pore diameter but the beam density in the PVF sample. [1] where P d = the porosity estimated by SEM observation in dry state, v = the volume in dry state, g = the weight of the volume in dry state, V = the volume in hydrated state Material Properties of PVF The material properties of the prepared PVF were investigated using the indentation tester shown in Fig.4. Each PVF test specimen (diameter: 60 mm; thickness: 5 mm) was soaked in phosphate buffered saline (PBS). A compressive load was applied via a 8-mmdiameter cylindrical probe, which was made from sintered copper balls with diameters of 0.5 mm. This probe was applied to the PVF specimen at a speed of 0.05 mm/s. When the [2] [3] Fig. 4. Schematic illustration of indentation tester and definitions of elastic index (= P/X) and viscosity index (=A). 251

4 maximum load of 1.0 kn was recorded, the probe was moved backward at the same speed. When the applied load decreased to 10 N, the compressive motion was reapplied. The elastic index (= P/X [Pa]) and viscosity index (= A [Pa]) were used to evaluate the material properties of the PVF specimens. Plastic deformation of the PVF has to be settled in the early stages for the practical use. Therefore, this load/unload cycle was repeated up to 500 times Tribological Properties of PVF Rubbing Motions and Wear Mechanisms of PVF As can be seen from the examples of tribological testing protocol of ASTM F [6], the pin-on-plate wear test is one of the most widespread and useful configuration tests. Despite the development and extensive use of simple pin-on-plate tests, the geometrical configuration between two bearing surfaces has to be very careful when the bearing material with low elastic modulus is used. In this case, there are clear differences in the contact stress between the pin and plate configuration, and misalignment between the pin and the plate may raise the stress within the tribological system [7]. A configuration of sphere-on-flat is of considerable benefit, because this point contact relieves the local peak stress raised by the misalignment. Relative movement between two bearing surfaces may be classified into 2 main groups, sliding and rolling. However, when there is a difference in severity of friction between two bearing surfaces, the sliding is sometimes classified into further 2 groups, sliding and gliding [8,9]. The kinematic conditions are represented graphically shown in Fig.5. In this study, the sliding or gliding motions were applied to PVF specimens to investigate the wear characteristics. Fig.5. Kinematic conditions of contact for the various test configurations: sliding, rolling and gliding, adapted from reference [8]. Three classical wear mechanisms have been established for the sliding or gliding of a hard counterface on a polymeric material such as a PVF. These are: (A) adhesive wear, (B) counterface or third-body abrasive wear and (C) fatigue wear derived from ploughing friction (Fig.6). In ploughing friction, which is generated when a hard bearing surface slides or glides on a soft bearing surface, the soft surface is deformed greatly, so that friction and wear is enhanced by ploughing. In this study, the gliding motions of a hard spherical surface on a PVF flat surface was adopted in order to investigate the fatigue wear of PVF specimens. A journal bearing is now the most popular tribological system in industrial filed, in which a sliding motion is predominating factor. Despite the extensive use of this cylinder column / hollow in design, the sliding direction has to be very careful when the bearing material with low elastic modulus is used. Over the past 20 years, laboratory studies on wear of polymeric materials have been conducted using unidirectional or reciprocating linear wear testing machines. However, unidirectional reciprocating motion causes the surface of polymers such as a polyethylene to become oriented and strain hardened. It appears that multidirectional motion causes the surface layer to be constantly redrawn and reoriented at acute angles. This leads to shearing of polymer particles from the surface, 252

5 Fig.6 Wear mechanisms. producing wear [10-12]. Therefore, in this study, the multidirectional sliding of a hard spherical surface on a PVF flat surface was also adopted in order to investigate the adhesive or abrasive wear of PVF specimens Tribological Properties of PVF (Gliding motions) The fatigue wear produced by ploughing friction under the gliding motions of the PVF specimens was investigated by using a reciprocating tester (Fig. 7). A 3-mm-thick flat plate of PVF was used as the lower specimen and it was attached to a liquid bath by a holder jig. The compliant stainless steel bearing was a sphere with a radius of 30 mm. PBS was used as the lubricating liquid. A load of 137 N, which was about 20% of a human body weight, was applied in order to elucidate the wear characteristic derived from the fatigue phenomena. A 20-mm stroke with a mean sliding speed of 20 mm/s, which was a simplified sliding motion in a natural synovial joint [13], was applied for the total sliding distance of 10 km, in which about 50% of the PVF specimens suffered from large ruptures. The damaged surfaces of the survived PVF specimens were observed by optical microscopy. Gravimetric wear of the PVF was also measured in order to investigate the tribological properties of the PVF specimens. Fig. 7. Schematic illustrations of ball-on-flat wear tester with reciprocating motion Tribological Properties of PVF (Multidirectional sliding motions) The wear characteristics in multidirectional slidings of the PVF specimens was investigated by using wear tester shown in Fig. 8. A 3-mm- thick flat PVF plate was used as the lower specimen and it was attached to a liquid bath by a holder jig. The compliant stainless steel bearing was a sphere with a radius of 30 mm. A load of 1.0 kn, which was approxmately 150% of a human body weight, was applied and PBS was used as the lubricating liquid. The sliding speed of the upper stainless ball was 72 deg/s, while that of 253

6 the lower PVF plate was 24 deg/s. The sliding speed between bearing surfaces was estimated about 40 mm/s, that was condidered to be encountered in a natural human joint. The total sliding distance was 10 km. The damaged surfaces of the survived PVF specimens were observed by optical microscopy and the apparent contact pressure on the PVF specimen during a test was calculated. Gravimetric wear of the PVF was also measured in order to investigated tribological properties of the PVF specimens. Fig. 8. Ball-on-flat wear tester with multidirectional sliding motion. Fig. 9. SEM images of various PVF samples prepared by varying the concentrations of the cross-linking agent (formaldehyde) and acid catalyst (sulfuric acid). 254

7 3. Results and Discussion 3.1. Polyvinyl formal (PVF) as a Bearing Material Figure 9 shows SEM images of various PVF samples formed by varying the concentrations of cross-linking agent (formaldehyde) and catalyst (sulfuric acid) used in the formalization reaction. Pore diameter, porosity, and beam density in the PVA specimens were shown in Figs Despite only PVA being used as the source material and the reaction temperature, reaction time, and other conditions being constant, a diverse range of structures were observed. Fig. 10. Pore diameter, porosity, and beam density in PVA specimens. Aqueous solution of PVA was 10.0 wt%. Fig. 11. Pore diameter, porosity, and beam density in PVA specimens. Aqueous solution of PVA was 15.0 wt%. 255

8 Figure 12 schematically shows the internal structures of the PVF specimens shown in Figs In Fig. 12, the number and size of the spheres indicate the number of pores and the average pore diameter per unit volume, respectively. The beam density is indicated by the degree of darkness of the shading. Such schematic diagrams are a very effective for visualizing the relationship between the material and tribological properties since they clearly show the effect of varying two parameters in the formation conditions on the internal structure of PVF. Hydration of the PVF did not affect the number of pores or the average pore diameter. However, hydration appears to affect the beam density between pores. Increasing the formaldehyde concentration increases the beam density between pores (e.g., A-D-G and L-O-R in Fig. 12), because the intermolecular and intramolecular reactions of PVA were accelerated. This accelerated cross-linking between PVAs reduces the flexibility of the beams, so that the change in the beam density on hydration is not obvious. The effect of sulfuric acid concentration on the internal structure of PVF is relatively small, because similar structures are observed at the same concentrations of sulfuric acid (e.g., A-B-C and P-Q-R in Fig. 12). Fig. 12. Schematic diagrams of the internal structures of PVF specimens in dry and hydrated states. 256

9 3.2. Material Properties of PVF Figure 13 shows the variation in the elastic index of PVF samples with the number of indentations. In almost all cases, the elastic index increases and finally becomes constant as the number of indentations increases. The PVF specimen whose elastic index did not become constant has a large number of pores per unit volume (e.g., A, B, J in Fig. 12). In the initial state, deformation initially occurs in a region with pores having larger diameters. In the static state, deformation of the pores is terminated and the stress of the PVF is elastically supported by the beams, so that the elastic index became constant. The elastic index after 500 cycles of indentation testing is shown in Fig. 14. The PVF specimens with large pore diameters and low beam densities (A, B, C, J, K, L in Fig. 12) have higher elastic indices. Fig. 13. Transitions of elastic index of PVF samples in accordance with numbers of indentations. 257

10 Fig. 14. Elastic index after 500 cycles of indentation testing. Fig. 15. Variation of viscosity index of PVF specimens with number of indentations. 258

11 This is due to most of the pores in these PVF specimens being easily crushed when a load is applied so that the load is supported by the crumpled mass of beams. By comparing the PVF specimens that have similar pore structures (E vs. F, N vs. O, M vs. P, H vs. I, and Q vs. R in Fig. 12), a higher beam density is considered to be responsible for the higher elastic index. The higher formaldehyde concentration is responsible for the higher elastic index for the 10 wt% PVA specimen. However, the 15 wt% PVA specimen exhibited the lowest elastic index when the formaldehyde concentration was 50 wt%. The beams of the PVF cross-linked in the specimen with a 50 wt% formaldehyde concentration may have a more flexible structure than that with a 100 wt% formaldehyde concentration, resulting in the lower elastic index. Figure 15 shows the variation in the viscosity index of PVF samples with the number of indentations. The viscosity index is thought to be affected by two factors: the viscosity of beam in the PVF and the fluid flow resistance of the PVF. Liquid migration in the PVF will occur when the PVF is deformed by an externally applied load. An increase in the flow resistance or a reduction in the pore size may suppress liquid migration, thereby increasing the viscosity index. Figure 16 shows the viscosity index after 500 cycles of indentation testing. In the aqueous solution of 10 wt% PVA, the viscosity index increased with an increase in the formaldehyde concentration. This confirms the results shown in Fig. 12, namely that a higher concentration of formaldehyde produces smaller diameter pores in the PVF specimens. Therefore, the fluid flow resistance increases, causing the viscosity index to also increase. The formation of a continuous film of lubricant between bearing surfaces may be attributed to conventional elastohydrodynamic lubrication, squeeze film effects, microhydrodynamic lubrication, and local deformation of the low elastic modulus layers. The PVF specimens prepared in this study not only had a low elastic modulus but also a highly porous structure, suggesting that their material properties will be optimal for a biomimetic bearing that mimics articular cartilage, exhibiting ultra-low friction and wear with high durability in a variety of lubrication modes [14]. Fig. 16. Viscosity index after 500 cycles of indentation testing Tribological Properties of PVF (Gliding motions) Figure 17 shows surface observations of the PVF specimens after ball-on-flat wear tests with reciprocating motions. The compliant stainless steel bearing surface was undamaged and was not scratched or marked in any way. This indicates that third-body abrasive wear did not occur in these tests. In each test, a depression developed in the PVF specimen under 259

12 Fig. 17. Optical microscope images of PVF samples after ball-on-flat wear tests with reciprocating motions. the compliant bearing. This depression is due to creep deformation of the PVF, which was assumed to occur in the indentation tests (Fig. 13). Large rupture behaviors could be classified into two main modes: wear mode (E and F) and plucking mode (A, B, C, D, G, J, K, L). As Fig. 12 shows, the beam densities of specimens E and F decreased on hydration. This weakens the beams, making the wear mode the principal mode. The PVF samples of A, B, C and J, K, L with high porosities and low beam densities could be easily plastically deformed, making the plucking mode the principal mode. It transpires that the absence of the acid catalyst (sulfuric acid) in a low PVA concentration solution deteriorated the wear characteristics (A, D, G vs. J, M, P). Figure 18 shows the gravimetric wear of PVF specimens after tribological testing. Although the beam densities of specimens M, N and O were lower than those of specimens P, Q and R, specimens M, N and O exhibited higher wear resistances. This result indicates that a suitable pores diameter, porosity and beam density of PVF can be achieved for developing a load bearing system with hydration lubrication. 260

13 Fig. 18. Gravimetric wear of PVF samples after ball-on-flat wear tests with reciprocating motions. Fig. 19. SEM images of wear debris generated from PVF samples after ball-on-flat wear tests with reciprocating motions. 261

14 Figure 19 shows SEM images of wear debris generated from PVF samples. The wear debris produced by specimens C, J, K and L appeared to be large fibrous aggregates. The wear debris produced by specimens that exhibited high wear-resistance properties was small, and in these samples the amount of wear debris remained at low Tribological Properties of PVF (Multidirectional sliding motions) Figure 20 shows optical microscope images of PVF samples after ball-on-flat wear tests with multidirectional sliding motions. Specimens O, Q and R showed high wear resistance; the other specimens, however, wore out prior to the end of the tests. The plucking mode in the wear process was observed in the PVF specimens A, B, C. On the other hand, specimens J, K, L, which have high porosities and low beam densities, were readily plastically deformed. Fig. 20. Optical microscope images of PVF samples after ball-on-flat wear tests with multidirectional sliding motions. 262

15 Figure 21 shows the apparent contact pressure on the PVF during a test and gravimetric wear of PVF samples after ball-on-flat wear tests with multidirectional sliding motions. According to the estimation of the contact pressure in a pig shoulder joint, apparent contact pressure was about 1.8 MPa at a load of 1.0 kn [14]. Although there is a different geometrical configuration in bearing surface between natural synovial joints and these multidirectional sliding tests, the frictional condition adopted in this test was thought to be approximately the same in a natural joint. The wear characteristics of the three specimens with high wear resistance for multidirectional sliding motions (O, Q and R in Fig. 21) were similar to those for reciprocating motions (O, Q and R in Fig. 18), although the degree of friction for multidirectional sliding motions was much higher than that for reciprocating motions. Fig. 21. Apparent contact pressure on the PVF during a test and gravimetric wear of PVF samples after ball-on-flat wear tests with multidirectional sliding motions. 4. Conclusion In order to produce the biomimetic bearing which is equivalent of the highly effective lubrication system in a human joint, the polyvinyl formal (PVF) with continuous porosity and hydrophilicity was proposed as a bearing material. The PVF was made by performing a chemical cross-linking reaction in polyvinyl alcohol (PVA), and its pore diameter, porosity, and beam density could be controlled by varying the concentrations of cross-linking agent (formaldehyde) and catalyst (sulfuric acid). It was clear that the PVF had high wear-resistance properties in comparison to the PVA hydrogel, and that a suitable pores diameter, porosity and beam density of PVF were thought to exist. Although the tribological property in fatigue wear produced by ploughing friction was inadequate, it was obvious that the PVF was a potential material for developing a load bearing system with hydration lubrication. References (1) Dowson, D. et al. Design considerations of cushion form bearing in artificial hip joints, Proc. Instn. Mech. Engrs., Vol. 205 Part H, 1991, p.59. (2) Auger, D.D. et al. Cushion form bearings for total knee joint replacement, Part 1: Design, friction and lubrication, Proc. Instn. Mech. Engrs., Vol. 209 Part H, 1995, p

16 (3) Oka, M. et al. Development of artificial articular cartilage, Proc. Instn. Mech. Engrs., Vol. 214 Part H, 2000, p.59. (4) Kobayashi, M. et al. Development of the shields for tendon injury repair using polyvinyl alcohol-hydrogel, J. Mioat., Vol. 58 Part H, 2001, p.334. (5) Nakanishi,Y. and Higaki,H., Japanese J. of Tribology, Vol.52, No.4, 2007, pp (6) ASTM: F Standard Practice for Reciprocating Pin-on-Flat Evaluation of Friction and Wear Properties of Polymeric Materials for Use in Total Joint Prosthese, Vol.12, 1991, pp , American Society for Testing Materials, Philadelphia, Pennsylvania. (7) A.A. Besong, Z.M. Jin, J. Fisher, Proc. Instn. Mech. Engrs., Part H, vol.215, 2001, pp (8) T.M. McGloughlin, A.G. Kavanagh, Proc. Instn. Mech. Engrs., Part H, vol.214, 2000, pp (9) G.B. Cornwall, J.T. Bryant, C.M. Hansson, Proc. Instn. Mech. Engrs., Part H, vol.215, 2001, pp (10) Bargdon,C.R., O Connor,D.O., Lowenstein,J.D., Jasty,M., Syniuta,W.D., Proc. Instn. Mech. Engrs., Part H, vol.210, 1996, pp (11) Joyce,T.J., Vandelli,C., Cartwright,T., Unsworth,A., Wear, vol.250, 2001, pp (12) Nakanishi,Y., Higaki,H., Umeno,T., Miura,H., Iwamoto,Y., Japanese J. Clin. Biomech., vol.28, 2007, pp (13) Higaki,H., Murakami,T., Nakanishi,Y., Miura,H., Mawatari,T., Iwamoto,Y., Instn. Mech. Engrs., Part H, vol.212, 1998, pp (14) Murakami,T. et al., Proc Instn. Mech. Engrs., Vol.212 Part H, 1998, p