The different stages leading to material loss are discussed in term of particle removal ..... to the last stage of degradation controlled by the de- bris behavior [12].
Journal of Sol-Gel Science and Technology 6,287-297 (1996) @ 1996 Kluwer Academic Publishers. Manufactured in The Netherlands.
Tribological Properties of Ormosil Coatings Laboratoire
Laboratoire
Laboratoire
I’. ETIENNE de Science des Materiaux Vitreux, Universite’ de Montpellier II, Place Eugene Bataillon, 34095 Montpellier Cedex j-France J. DENAPE AND J.Y. PARIS de Genie de Production, Ecole Nationale d’Ingt%ieurs de Tarbes, Av. d’Azereix, BP 1629, 65016 Tarbes Cedex-France J. PHALIPPOU AND R. SEMPERE de Science des Mate’riaux Vitreux, Universite’ de Montpellier II, Place Eugene Bataillon, 34095 Montpellier Cedex j-France Received March 28, 1995; Accepted February 5, 1996
Abstract. Transparent plastics are not scratch resistant. The damage leads to a loss of optical properties. Coatings prepared using either tetraethoxysilane or colloidal silica particles embedded in “glymo” is a way to avoid these disadvantages. Tribological experiments are carried out to better understand the surface modification due to a sliding friction. It is shown that the wear track is not directly related to usual mechanical properties such as Young’s modulus and the hardness of the coating. The different stages leading to material loss are discussed in term of particle removal and debris circulation (accumulation or elimination) through the friction track. The mechanical properties of the film combined with the film to substrate adhesion are expected to play an important role as it can be deduced from results obtained as a function of the coating composition. Keywords:
coating on plastics, composite film, coaling tribology
1. Introduction Transparent plastics are promising materials for optical application. They are of a low cost and their density is well lower than that of usual glasses. Moreover, plastic glasses do not break easily. Such qualities offer a lot of industrial applications in the field of ophthalmicglasses. However they are not scratch resistant enough and then they lose very quickly their optical properties due to surface damage [ 11. Coating of these plastics with a hard material seems to be a means to overcome the observed disadvantages. This goal was achieved usiing both polysiloxanes and colloidal particles of silica [2,3]. Films of thickness varying between 2 and 7 pm, are easily deposited by the dipping or spinning technique [4].
It was previously demonstrated that the mechanical properties of the films depend on the nature of the silica re-inforcement as well as the volume percent of the organic compound that undergoes the polymerisation reaction. It was particularly shown that the Young’s modulus (obtained by a 3 point bending [5] or by nano indentation experiment [4]) is higher for coating prepared with a polysiloxane compound. Ultramicrohardness experiments lead to the same conclusion. However there is not a straightforward relationship between the elastic properties or the hardness and its scratch strength. In this paper, we investigate the tribological properties of transparent plastics coated with both systems (polysiloxane and colloidal silica). A coating is often required to achieve a low friction coefficient [6], but
288
Etienne et al.
coatings which decrease friction often increase wear [7]. The tribological behaviour of such materials can be successfully interpreted using the “third body approach” [S]. This theory is based on the idea that a dynamic screen is formed. This screen is built by the debris particles running through the surfaces in contact. The accumulation of particles between the contacting bodies builds a bed of powder leading to a separation of the surfaces previously in contact. This phenomenon involves a mechanism of load carrying capacity as occurs in hydrodynamic lubrication and for which wear is weak. On the other hand, the elimination of debris out of the contact enhances the interactions between the two solids and wear increases. Moreover, friction forces increase when the amount of particles in the contact increases [9]. Our work focuses mainly on damage and wear track and little attention will be paid to friction.
2.
Experimental
Procedure
The transparent plastic used as substrate for this study is a poly-diethylene-glycol-dialylbicarbonate. It is well known for its optical properties. This substrate is obtained by low temperature polymerisation. Young’s modulus of such a substrate is 2.3 GPa and its flexUral strength is 45 MPa. The solution used to coat the substrate contains y-glycidoxypropyltrimethoxysilane, a chemical compound usually named “glymo”. To this base compound is added an amount of silica. There are mainly two ways to provide this addition: tetraethoxysilane (TEOS) or an alcoholic solution of colloidal silica, which are used in this work. The behavior of this solution is a quite different from that of usual organometallic solutions. Hence a coating thickness in the range of 5 pm may easily be obtained using Table 1.
List of chemical
TEOS
+ Ha
Figure
1.
sol.
IO.lNI
‘plymo”
+ Hd
,d.
Details of the experimental
lO.lNI
Cdlddd
Source
0
y-glycidoxypropyltrimethoxysilane Tetraethoxysilane Colloidal
silica in methanol
dcohdk
ad.1
classical dipping technique. Details of experimental preparation are given in Fig. 1. Table 1 lists more information on the cited chemical compounds. The coatings prepared using “glymo” and TEOS are labelled as Te series and those prepared using “glymo” and colloidal silica as Re series. The coatings have been previously characterized [5]. Young’s modulus as well as the hardness of these films increase with the silica content. The tribological experiments are performed on a pinon-disk rig associated with a computer recorder. The device works with a linear reciprocating motion. It uses a hemispheric pin (12.5 mm radius of curvature) made ofpolished lOOG6 steel (E 52100 steel from SAE-AISI standard). The hardness of the pin is about 8 GPa and its elastic modulus is 210 GPa. The moving amplitude of the pin is 5.3 mm and the average speed is 0.05 m/s. The applied force is kept constant of 5 N.
Formula
Diethylene-glycol-dialyl-bicarbonate
II”
preparation.
compounds.
Compound
.Nca
O(CH*-cH&LcH*-cH=cH*)~
CR39
/O\
. CH~VCH~-CH~-(CH~)~-S~(OCH~)J Si(OCzHd4 SiOz particules,
20 nm diameter
Fluka,
Assay >97%
Fluka,
Assay >98%
NISSAN Sun Colloide MAST Type
Tribological Properties of Ormosil Coatings
The experiments are performed over a short time (less than 6 minutes), the whole distance of abrasion never exceeds 15 m. All the experiments are performed in ambient air. The room air moisture was not controlled but, in plastics, stress corrosion effects due to water may be likely ignored [ 101. The coefficient of friction is recorded continuously by measuring the friction force as a function of time. The coefficient of friction (p) is defined by the ratio between the tangent force F due to friction and the applied normal force M : p = F/M. The abrasion degree is evaluated by the volume of the removed matter. Optical and SEM observations correlated with tridimensional profilometry measurements are used to examine the wear track and the effects of the abrasion. The radius of curvature of the profilometer stylus is 5 pm with a vertical resolution of about 0.1 pm. Cracks are detected but their real depth cannot be evaluated. 3. 3.1.
Experimental
Results
Friction and Wear
qf the Bare Substrate
The friction response is quite sensitive to the sliding rig and the amount of debris trapped between the pin and the sample. The linear reciprocating tribometer gives high coefficient of friction (close to 1). This unexpected result is related to the reciprocating technique. Debris is easily recovered and trapped between the contacting bodies. In such a case it plays the role of an abrasive material. The friction coefficient is then higher than expected. The wear track has been observed after each experiment. The runs have been intentionally limited to five. The wear track shows a lateral extent of about 800 pm. It can be divided in two regions (Fig. 2a). The core of the track shows a coarse surface which originates from a high particle removal and an accumulation of debris which strongly sticks to the wear surface. Both sides of this central track exhibit an array of cracks (Fig. 2b). This surface damage interacts with the visible incident light and then induces light scattering. The goal of a coating is obviously to decrease this scattering effect. 3.2.
Friction and Wear Experiments on Coated Substrate
The friction coefficients of the coated plastics are higher compared to the bare substrate (Fig. 3). This
289
Table 2. Wear rate of coatings as a function of the silica content. The wear rate unit is wm/min according to the observation that tracks have identical length and width.
Compound
Silica (wt%)
Wear rate (&min)
Comments
Substrate
0
0.5
Re series
10
0.3
Semtctrcular cracks and abrasive grooves
30
0.2
Semicircular and lateral cracks
50
0.1
Partial removal of the coating
10 to 40
Highest values
Te series
Complete removal of the coating
unexpected effect appears clearly for the low silica content coatings. However the coefficient of friction decreases as the silica content in the film increases. The wear of coated substrates shows very important differences depending on the type of coating and the silica content (Table 2). a) Coatings Prepared Using “glymo” and TEOS (Te Series). Whatever the amount of silica, the extent of the cracking pattern on both sides of the wear track is less pronounced for the coated samples than for the bare substrate. The lateral cracks have a higher density but they do not extend very far from the wear track or they do not emerge at the surface. The material removal affects the central area of the track (350 pm wide) as previously mentioned (Fig. 4a). The surface damage occurs by removal of the coating. On both sides of the track, the coating is broken in large pieces and a sharp step is seen at this location (Fig. 4b). The step height is 5 pm which corresponds to the coating thickness. The initial size of the debris coming from the coating increases as the silica content increases (that is when the “glymo” quantity decreases). In the vicinity of the track filled sides, no wear occurs on the substrate after removal of the coating (Fig. 5). Only debris filling the cracks are observed. However, in the middle of the track, the wear process also affects the substrate and the debris consists now of a mixture of an organic compound with an organo-mineral compound which strongly sticks to the wear track. A schematic picture of the wear effect can be drawn (Fig. 6) according to the anaylsis performed by tridimensional profilometry.
290
Etienne et al.
(a)
(b) Figure 2. Wear track of the uncoated substrate showing both abrasion and lateral cracking view and (b) details of the crack pattern on the track sides (SEM observations).
using the rectprocating
linear tribometer.
(a) Genera
Tribological Properties of Ormosil Coatings
291
significant material loss is only observed for the highest silica content. 4. \
TE series
0,95 0
10
20 Silica
30
40
50
content (in weigth percent)
Figure 3 Variation of the friction coefficient of the coated substrates as a function of the silica content (a) for the Te series (“glymo” and TEOS) and (b) for the Re series (“glymo” and colloidal silica).
The free debris observed in the contact zone consists of a fine powder which fills up the cracks and accumulates against the lateral steps on the both sides of the track. Initial removal of large lumps of coating is not observed. b) Coatings Prepared Using “glymo” and Colloidal Silica (Re Series). The samples using Re coatings show quite different behaviour compared to Te coatings. The coating damage seems very homogeneous with a higher resistance to material loss. The samples having a low silica content (10 weight percent) clearly show a rather high quantity of semicircular cracks and abrasive grooves on the overall track (Fig. 7). Cracks interaction does not lead to large debris. When the silica content increases (30 weight percent) the wear track is more complex. Two crack patterns can be distinguished. The core of the track shows similar semicircular cracks as observed on samples with low silica content. These cracks are not as visible as the previous ones and do not extend on long distances (Fig. 8a). On both sides of the core track, lateral cracks associated to plastic grooves are observed. They are parallel and the space between two successive cracks is very small (Fig. 8b). The samples having the highest silica content (50 weight percent) show the first significant removal of material in the central region of the track (Fig. 9a). Large wear debris may appear at the onset of the sliding (Fig. 9b). The wear depth also reaches 5 pm whcih corresponds to the coating thickness. In all cases, the contact pressure leads to plastic flow (creep) of both coating and substrate. The residual vertical deflexion is 3 pm after running (Fig. 10). Cracking occurs regardless of the silica content but a
Discussion
The different damage patterns described above are representative of the same general wear mechanism taken at different stages of its progress. Three different stages can be discussed in terms of debris behaviour in the contact [ 111: damage initiation without material loss, particles separation and debris circulation (accumulation or elimination) through the friction track. Stresses generated during the friction, compared with the film to substrate adhesion strength, play an important role as they favour or delay the transition to the last stage of degradation controlled by the debris behavior [12]. These stresses are modelled as a combination of a residual stress present in the coating, an indentation stress field and a frictional stress field [13]. Adhesion is known to depend on the coupling agent concentration [ 141. However, for high concentrations (> 10 wt%), adhesion does not increases anymore. With 50 to 90 wt% of “glymo”, adhesion is probably independent of our coatings composition. Residual stresses can appear during the polymerization according to the shrinkage of the coating. We performed curvature measurement on a one side covered substrate to evaluate the residual stresses. The low curvature observed shows that the coating undergoes very slight tensile stresses. They seems quite independent of the silica nature and its content. Indentation stresses are known to increase with the hardness to Young’s modulus ratio (WE) of the coating [ 131. Previous results show that the I-I/E ratio of our coatings increases with the silica content [4]. So, higher value of silica content leads to higher intentation stresses. Moreover, the coating damage which may occur when a rigid coating is deposited on a soft substrate [ 151 is not expected in our material because both coating and substrate exhibit moduli in the same range. Frictional stresses are functions of the load and the coefficient of friction. As the load is constant, Fig. 3 suggests that these stresses should decrease when the silica content increases. Assuming adhesion at the coating-substrate interface and initial internal stresses are constant, the stress involved during the experiment should be the result of a competition between indentation and frictional stresses. SEM observations show that when the silica
292
Etienne et al.
(a)
(b) Figure 4. (a) General view of the wear track observed on the Te coated substrate (40 weight percent) showing the coating removal of the track sides and the crack pattern affecting the coating and the substrate floor (SEM observations).
and (b) details
Tribological Properties of OrmosiI Coatings
Figure 5. Tridimensional view of the wear damage recorded by profilometry (surface size: 500 x 500 wm2) on the Te coated substrate (40 weight percent).
free debris trapped against the steps (protecclve screen by load carrying elfect)
debris grimed on tha abraded substrate (zone of highest press”~ of contact)
Figure 6. Schematic picture of the wear profile showing the coating damage drawn from profilometry data.
content increases, larger debris is removed. Consequently, the indentation stress increase is not balanced by the frictional stress lowering.
With a reciprocating displacementof the pin, the net result is the observationof the interaction of two arrays of cracksoriented accordingto the two oppositesliding directions. The cracks are theoretically equally spaced. The space between two adjacent cracks is related to the friction coefficient [ 181. It is small if the friction coefficient is high. On the other hand, the length and the depthof the cracksareclosely related to the mechanical propertiesof the material (hardness,Young’s modulus and fracture toughness)[ 191. The secondcrack pattern is observed for coatings having an intermediate silica content. It is usually encounteredon brittle materialssubmittedto scratchtests [20]. It refers to a surface subjected to a high local pressurewhich can involve a limited plastic flow from where cracking occurs (sharpcontact characterized by small radius of curvature). Such cracks are usually straight and parallel, symmetrical against the central plastic groove and oriented towardsthe oppositedirection of the stylus displacement[21]. Local abrasion grooves are also observed on the coated samplesand may be related to a material transfer adhesiveonto the metallic pin. This first stageof crack formation occurs on all the samples.It is still observedon both sidesof the wear tracks (where the contact pressureis lower) on the Te samples,as well ason the bare substrate.
b) Particle Detachment. a) Damage Initiation
by Cracking. This first stage is
clearly observed on the Re coatings with a low silica content. No significant material loss is observed and only semicircular cracks are seen, Such cracks involve an elastic response of the sample (Hertzian contact characterized by large radius of curvature) and are usually found with brittle materials like glasses and ceramics [16]. They originate from the stress distribution in the material that is induced by the pin displacement. In front of the contact a compressivestress appearswhich is balanced by a tensile stressbehind the pin [ 171. Crack formation occurs when the elastic energy stored in the material is transformed into surface energy which indeed allows stressrelease. After the first crack appearance,the elastic energy increases again to level correspondingto the formation of a new crack. Thesecrackshappenin the tensile zone (modeI) behind the pin and show a half-ring shapewhosecurvature is oriented towards the direction of the pin motion.
293
The secondstagestarts with the removal of large particles of coating as the result of the intensive prior cracking of the surface. The first particles are detached from the middle of the friction track where the pressureis the highest. The following runs widen the wear track until the contact pressure becomeslow enough to avoid further damage. Then, the removed debrisare finely ground inside the contact until the surfaceenergy which is createdis balancedby the associatedfracture energy [22]. The accumulation of such moving and rolling fine particles builds a bed of powder which leads to the separation of the two contacting bodies [23]. This protective screenof free debrisallows the velocity accommodationbetweenthe contacting surfacesby a load carrying effect analogous to that observedin lubrication with a fluid [24]. This stageis characterized by an increaseof the coefficient of friction showinga higher sliding resistancedueto the rise of a great amount of particles betweenthe sliding surfaces. The wear profile of the track has U shape whosedepth correspondsto the coating thickness.
294
Etienne et al.
(a)
(b)
V
Figure 7. Aspect of the wear track observed on the Re coated substrate having a low silica content (10 weight percent). (a) General view of the friction track showing the semicircular crack pattern (SEM observation) and (b) details of the crack pattern recorded by tridimensional profilometry (surface size: 750 x 100 pm2).
c) Debris Circulation.
The debris behaviour (composition andlocation in the contact) plays a dominantpart in the third stageof damage. The nature of debris evolves because,under the action of the sliding friction, a chemical reaction may occur. Pure silica colloidal particles can react with air moisture to give rise to hydrated species [25]. Furthermore humidity stimulates the fracture of oxide ceramics[26] and contributesto achievea very fine size of debris.
Due to the hemisphericshapeof the pin, the ground particles are drained off the borders of the wear track where the pressureis lower. A part is ejected out of the contact and no longer participates in the load carrying mechanism.The other part cannot escapeand piles up againstthe steepsidesandthe external floor of the wear track where they act as a protective screen. In the centre of the track where the pressureis the highest, the particles are not numerousenough to assumetheir role of screen. On the contrary, they are trapped in
Tribological Properties of Ormosil Coatings
295
(4
(b) Figure 8. Aspect of the wear track observed on the Re coated substrate having an intermediate silica content (30 weight view of the friction track and (b) detail of the lateral cracks associated with a plastic groove (SEM observations).
percent)
(a) General
296
Etienne et al.
(4
Figure 9. Aspect of the wear track observed on the Re coated substrate having the highest silica content (50 weight percent). (a) General view of the friction track showing the crack patterns and the coating removal (SEM observations) and (b) details of the coating damage by removal of a large particle detached by cracks interactions (tridimensional profilometry, surface size: 600 x 400 pm2).
Figure 10. Deflexion of both coating and substrate under the contact pressure (tridimensional profilometry, surface size: 1500 x 200 pm’). Wear affects only the lowest zone of the friction track (Re coated substrate with 50 weight percent silica content).
Tribological Properties of Ormosil Coatings
the track, and they stick to the sliding surfaces and contribute to wear damage by abrading the opposite material. This stage is characterized by a progressive increase of the wear depth in the center of the track (aggressive role of the debris) while the external floor of the wear track remains flat and undergoes no further damage (protective role of the debris). Hence, the wear behaviour of coated substrates is not straightforwardly related to the mechanical properties such as Young’s modulus and the hardness, of the coating. As an example, coatings prepared using TEOS (Te series), which exhibit the highest elastic modulus and hardness, are more easily damaged than coatings prepared with colloidal silica filler. However, experiments must be done to evaluate the adhesion and the fracture toughness as a function of their composition. 5.
Conclusion
The tribologic properties of optical polymer may be improved by coating them with a modified silica film using a dipping technique. The surface damage exhibits cracking and coating removal by abrasion depending on the silica content of the film. In this respect, films prepared using a colloidal silica solution exhibit the best tribologic properties. Even in that case, it appears necessary to optimise the composition. With a low silica content the crack expansion is severe. With too high a silica content the indentation stress is too large and the coating may be removed by friction. Three successive stages of surface degradation have been distinguished: damage initiation by cracking, particle detachment, and debris circulation (accumulation or elimination). The adhesion strength compared to the residual, indentation and frictional stresses play a major role during the two first stages. The last stage of degradation is mainly related to the debris behaviour which acts as a protective screen or an abrasive body
297
depending on the location of the debris in the contact area. Acknowledgments The authors gratefully acknowledge the continuing aid of the ESSILOR Research Department. References 1. BJ Briscoe and D. Tabor, Brit. Polymer, J. lo,54 (1978). 2. J. Hennig. Kunststoffe 71, 103 (1981). and K. Piana, Mat. 3. K. Greiwe. W. Glaubitt, S. Amberg-Schwab, Res. Sot. Symp. Proc. 271,725 (1992). of Montpellier II (1993) 4. P. Etienne, Thesis, University 5. P Etienne, R. Sempere, and J. Phalippou, J. Sol-Gel Science and Technology 2, 17 1 (1994). 6. J. Halling, Thin Solid Ftlms 108, 104 (1983) ASLE Trans. 10. 1 (1967). 7. E. Rabinowicz, 8. M. Godet, Wear 136,29 (1990). 9. J. Denape and J. Lamon, J. Mat. Sci. 25.3592 (1990). 10 J.G. Williams, Fracture Mechamcs of Polymers (Wiley, NY, 1987) p. 189. 11. C. Colombie, Y. Berthier, A. Floquet, L. Vincent, and M. Godet, J. of Tribology 106, 194 (1984). 12 S J. Bull, Surf. Coat. Tech. 50, 25 (1991). 13. PJ. Burnett and D S. Rickerby, Thin Solid Films 157, 233 (1988). 14. P. Walker, Silanes and Other Coupling Agents, edited by K.L. Mmal (VSP, 1992). p. 21. 15. PM. Ramsey, H.W. Chandler, and T.F Page, Surf. Coat. Tech. 49.504 (1991). 16. D. Tabor, J. Lubri. Techn. 103, 169 (1981). 17. B.R. Lawn and R.T. Wilshaw, J. Mat. Sci. 10, 1049 (1975). 18. M.V. Swain. Fracture Mechanims oj Ceramics, edited by R.C. Bradt, D.PH Hasselman and EE Lange (Plenum Press, N.Y. 1978) Vol. 3, p. 257. 19. D B Marshall, Progress in nitrogen ceramics, Nato-asi serie E. Applied Sciences 65,635 (1983). 20. J.C. Conway Jr. and H.P. Kirchner, J. Mat. Sci. 15.2879 (1980) 21 D.H. Buckley and K. My~oshi, Wear 100,333 (1984). 22. 0.0. Ajayio and K.C. Ludema. Wear 140, 191 (1990). 23. M. Godet, Wear 100,437 (1984). 24. Y. Berthier, Wear 139, 77 (1990). 25 T.E. Ftscher, Ann. Rev. Mater, Sci. 18, 303 (1988). 26. T.A. Michalske and B C. Bunker, J. Appl. Phys. 56,2686 (1984).
