J Sol-Gel Sci Technol DOI 10.1007/s10971-015-3781-6

REVIEW PAPER

Mechanical properties of sol–gel coatings on polycarbonate: a review Nicolas Le Bail1,2 • Ste´phane Benayoun1 • Be´range`re Toury2

Received: 21 May 2015 / Accepted: 22 June 2015 Ó Springer Science+Business Media New York 2015

Abstract Organic/inorganic hybrid coatings prepared via the sol–gel process have received lot of interests during the early twenty-first century. Devices obtained thanks to this low-temperature route display a large panel of bulk and surface properties that can be modulated according to the target. Moreover, this versatility enables to offer solutions in various domains and industrial applications such as microelectronics, optic, aeronautic, automotive, health. When the aimed application required polymer as substrate, the use of sol–gel process takes its full interest as soft chemistry. This review is dedicated to mechanical properties improvement of a common polymer substrate, e.g., polycarbonate, when a transparent sol–gel coating recovered it. Graphical Abstract

& Be´range`re Toury [email protected] 1

Laboratoire de Tribologie et Dynamique des Syste`mes, Ecole Centrale de Lyon, 69130 Ecully, France

2

Laboratoire des Multimate´riaux et Interfaces, UMR 5615, Universite´ de Lyon, 69622 Villeurbanne, France

Keywords Sol–gel  Coating  Polycarbonate  Mechanical properties

1 Introduction Polycarbonate (PC) is a high-performance amorphous engineering thermoplastic used in widespread applications; its good mechanical properties and transparency make it a good candidate for such a low price. Nevertheless, it displays two main limits: The first is linked to its low resistance to scratch and abrasion, and the second concerns its vulnerability to UV, hot stream and certain solvents and chemicals [1]. The latter occurring in the longer term, polycarbonate mechanical enhancement is more challenging to further increase its durability. In this way, two solutions can be investigated: a protective coating deposit or a bulk modification [2, 3]. This review is only focused on the first solution and more precisely on silica-based coatings obtained by the sol–gel process, allowing fulfilling all the required criteria (in terms of implementation facility and performances). Actually, many properties can be reached for such coated device: anti-reflection, flame retardant [4], UV protection [5], self-cleaning [6], etc. As said above, such new functionalities become interesting only if mechanical properties of the whole device can be enhanced. Presently, this has already been proven by results reported in the literature; for example, a basic sol– gel coating based on classical silica-based precursors (GPTMS: (3-glycidoxypropyl)methyltrimethoxysilane, TEOS: tetraethoxysilane) can reach a hardness of 71 ± 25 HV [7] and a Young modulus of 10 GPa [8, 9] when PC hardness is only 13 ± 2 HV [7, 10] and Young’s modulus is around 3 GPa [9].

123

J Sol-Gel Sci Technol

Pure silica coatings are obtained through hydrolysis– condensation reactions (Fig. 1) of basic alkoxysilane, such as TEOS, precursor appreciated for releasing only ethanol during hydrolysis. Moreover, the use of a pure inorganic network is strongly limited in terms of implementation and new functionality input. A way to facilitate the shaping step and to tailor both bulk and surface properties of the final coating is to bring an organic component covalently linked to the network through the use of an organoalkoxysilane [R0 (4-x)–Si(OR)x]. In addition to bringing flexibility to the final network, R0 , the organic part plays a key role to modulate either chemical or mechanical properties. Organic components can be introduced into the inorganic network in two different ways as network modifiers or network formers [12]. Network modifiers contribute to the functionalization of the matrix. For example, methyltriethoxysilane (MTES) can give flexibility [13] when octyltriethoxysilane (OTES) or fluoro(triethyl)silane (FTES) is mostly known to bring hydrophobicity [14, 15] to the final device. More exotic molecules are reviewed elsewhere [12]. Concerning the network formers, 3(trimethoxysilyl)propyl-methacrylate (MEMO) [16–18], (3-aminopropyl)triethoxysilane (APTES) [19] and (3-glycidoxypropyl)methyltriethoxysilane (GPTES) [20, 21] are widely used. These precursors, which have to be thermally or photo-chemically cured leads to an organic network, tangled up to the inorganic one. Close to this approach, it is also reported that the addition of pure organic monomers to the network former performs a double reticulation and creates a convolution of both O/I networks [4, 10, 22–25]. In any case, sol–gel process mastery is primordial to prepare a stable sol, able to be deposited and reticulated in a three dimensions O/I hybrid network. With the aim to keep PC aesthetical values, the coating transparency becomes an additional feature to consider. Hence, it is very important to master the sol properties in order to avoid any phase separation [10], precipitation or sol trouble. Hydrolysis/condensation sol–gel reactions have to be controlled as well as the pH, the hydrolysis factor and the solvent ratio. All those parameters have an impact on the organization of the inorganic network. Actually, it is a low-temperature process and coating after curing can enhance surface properties of the PC.

Hydrolysis: Si-OR + H2O Si-OH + R-OH Condensation: Si-OH + HO-Si Si-O-Si + H2O Si-OH + RO-Si Si-O-Si + R-OH Fig. 1 Hydrolysis/condensation reactions obtained by the sol–gel process on silica-based precursors [11]

123

This review reports all strategies retained to enhance mechanical properties of polycarbonate when a sol–gel coating is deposited. The main objective is to evaluate the influence of synthesis (organic content, filler additivation, etc) and process parameters (thickness, aging time, curing process, etc) on the modification of the film mechanical performances. At the same time, we discuss on how the mechanical parameters (hardness, Young modulus, critical load) affect the abrasion or scratch resistance. However, prior to that, some mechanical models are basically explained.

2 Mechanical properties, basic models The objective here is not to describe in detail the entire mechanical models existing but to make an overview and explain some mechanical calculation basis dedicated to coatings.

2.1 Indentation Indentation tests permit to characterize the plastic deformation of a material by applying a compressive force F with a tip (usually made of diamond or WC–Co). The material behavior is characterized by an indentation pressure which is estimated by the ratio (applied force)/ (residual print surface). Brinell’s (with a steel sphere with R radius), Vickers (diamond pyramid square based, angle between opposite faces: 136°) and Rockwell’s (diamond cone or steel sphere) methods permit to measure the material hardness. Brinell’s (HB) and Vickers (HV) models are defined as: HB ¼

2P pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pD0 D0  D02  d 02

ð1Þ

P ½26 d2

ð2Þ



HV ¼ 1:854 

HB Brinell hardness, HV Vickers hardness, P applied load (kg), D0 indenter diameter (mm), d indentation diagonal (mm), d0 indentation diameter (mm). Rockwell’s method is based on the determination of the permanent penetration depth when Vickers and Brinell’s one are based on the evaluation of the residual diagonal and diameter, respectively. Finally, those methods are accepted as soon coating thickness (e) is ten times superior to the penetration depth [26]. Another point which creates problems is the load range used; the indentation size effect induces a dependence of the hardness with low loads (P \ 5 N typically) (all the details are explained elsewhere) [27].

J Sol-Gel Sci Technol

Another important parameter to control is the elastic moduli which measure the resistance to deformation of materials when a force is applied. It is defined, for isotropic material (it is the case of polycarbonate), as the ratio between the applied uniaxial stress (r) and the resulting strain (e) in the same direction [28]. Nanoindentation technique takes its origin in the previous models, but only small penetrations under the micron are performed and the displacement coupled with the load applied is continuously recorded (example in Fig. 2). The small penetration depth and the contact stiffness (indentor material, see Fig. 2) accessible with this kind of equipment can permit to determine the hardness and the elastic moduli of the material [29–31]. The first indentation device was introduced by Martens [32] and developed by Pethica et al. [33] in the 1980s. It has to be mentioned that it is very complex to evaluate the stress field under the tip, and as a result, the data interpretation can be very tough. The most used indenter is a Berkovich tip (three side pyramid), and the analysis to evaluate the elastic modulus and hardness is explained by the method of Oliver and Pharr [29] which develops the ideas of Loubet et al. [34] and Doerner and Nix [30]. The projected contact area A has to be determined and a calculation can be made, assuming the Young’s modulus is independent of depth [29]. Therefore, hardness and Young’s modulus follow from: H¼

Pmax A

ð3Þ

And with a polycarbonate substrate 1 1  t2 1  t02 1  t2 þ ¼  Eeff E E0 E

ð4Þ

where Eeff

pffiffiffi S p ¼ pffiffiffi 2b A

ð5Þ

Eeff is the effective elastic modulus, and t is the Poisson’s coefficient. E0 and t0 are dedicated to the indenter [E0 = 1140 GPa and t0 = 0.07 for a diamond indenter (ISO 14577)], and b is a constant (b = 1.034 for the Berkovich). In the case of Vickers or Brinell’s experiments, if we neglect the elastic component of the strain (the contact area during the indentation is equal to the surface of the print), we notice that the expressions of the hardness with those of the nanoindentation are equivalent if a geometrical coefficient is introduced. So the relation between the hardness in nanoindentation, H (MPa), and the Vickers hardness HV is: H ðMPaÞ ¼ 9:81ð2=1:854ÞHV  10:58 HV

ð6Þ

For the determination of the hardness, we have already underlined that the criterion of homogeneity of the deformed zone of the material is of 10h, with h the penetration depth. For the determination of the Young’s modulus, this value is higher because the field of the elastic strain extends far beyond 10h. So it is necessary to determine the value of H and E of the coating to use models which take into account the influence of the substrate even in nanoindentation. We shall quote in particular the references [27, 35–37] which review this problem. 2.2 Scratch resistance The scratch resistance is typically characterized by microscratch test. The principle consists of moving the indent on the surface while the applied force is constant or increases. The indenter used in scratch test is a Rockwell tip, different tip radius can scratch the surface, commonly comprise between 50 and 800 lm. The scratch mark is then analyzed by microscopy to reveal where failures are. Different kinds of defect can be described [38–40]: LC1 is the critical load where the first crack appears, and LC2 is the one where delamination and spalling is observed (Fig. 3). Other parameters can be determined like the groove width or the evolution profile of the tangential force recorded. Finally, it is also possible to have information about the adherence; this point will be discussed in Sect. 2.4. 2.3 Wear and abrasion resistance

Fig. 2 Nanoindentation data, load–unload curves with S = dP–dh (contact stiffness)

The abrasion test is often realized with a Taber. It consists of abrading a surface with a chosen abradant at a chosen force. Then, in case of transparent systems, the surface is often characterized in transmittance to evaluate the loss of transparency. Hardness, elastic modulus and critical loads are the most used parameters to characterize coatings, and one of the major characteristic searched (with scratch resistance) is a wear-resistant coating. Generally, the behavior in the

123

J Sol-Gel Sci Technol

LC1

LC2

Force (N)

Fig. 3 LC1 and LC2 examples observed on a sol–gel coating onto polycarbonate

abrasive wear is interpreted with regard to the Archars’s law: V ¼ ðKLFn Þ=H

ð7Þ

where V is the total volume of wear debris produced, K is a dimensionless constant, Fn is the total normal load, L is the sliding distance, and H is the hardness of the softest contacting surfaces. However, this expression introduces an experimental parameter, K, widely discussed for its amplitude of very large variation (10-1–10-8) [41, 42]. Contrary to what it was primary thought, it has to be mentioned that it is not necessary to have the highest Young’s modulus and hardness to obtain a wear-resistant coating [43, 44]. A solution proposed is to generate a large H/E ratio (‘‘elastic strain to failure’’ for a brittle material) [45, 46]. Indeed, a large H/E ratio induces a reduced contact pressure, and it is thus linked to a high elastic strain rather than a plastic deformation. Moreover, investigations were made on the relation between the coating’s resistance to plastic deformation and the H3/E2 ratio [47–49]. High value of this ratio indicates, when there is contact, a highly elastic behavior, whereas a low one reveals a plastic behavior of the film. 2.4 Adherence Adherence can be evaluated by crosscut test (ISO 2409). The principle consists of crisscrossing the coating surface with a razor blade by forming 25 squares (Fig. 4). Normed adhesive is then stuck and pulled off. Finally, a microscope analysis reveals whether the coating remained on the substrate or not. Adhesion is measured by calculating the percentage of coating pulled off ranging from 0 (excellent with 0 %) to 5 (very poor with [65 %) [6].

Fig. 4 Resulting image after crisscrossing the coating, each square measures 0.1 9 0.1 cm2

123

A way to estimate precisely the adhesion is to use of a Delaminator Test System, and this is double cantilever beam (DCB) technique. It permits to measure the adhesion in J.m-2 (Gc) by applying loads/unloads at a controlled displacement rate to produce coating cracks growth [50]. Gc is then calculated using the Kanninen corrected equation [51]: a 2 12  P2c þ 0:64 Gc ¼  ð8Þ h E  B2  h where Pc is the load where a crack is initiated, E is the Young’s modulus of the substrate, B is the substrate width, a is the crack length, and h is the substrate thickness. Finally, it is also possible to qualify the adherence by measuring a Gc0 with the scratch test technique. It has to be mentioned that this Gc0 is not comparable to the Gc calculated from DCB. Indeed, the evaluation of this parameter is model dependent. Nevertheless, it is also a surface unit energy associated with crack propagation at the interface. Le Houe´rou et al. [52] propose a model to evaluate the adhesion thanks to the blister shape obtained by scratch test, and the method is explained in detail elsewhere.

3 Synthesis parameters influence 3.1 Organic content As explained in the ‘‘Introduction’’, there are two ways to introduce organic part within the inorganic network: the use of mixed precursor and the addition of polymerizable organic monomers. First, the effect of the use of an organoalkoxysilane [R0 of R0 (4-x)–Si(OR)x] tailored to bring an organic part is discussed. It is reported by Mizuta et al., who work on a system based on triethoxyphenylsilane (PhTES) and TEOS, that while the organic content increases within the coating; pencil hardness tends to be slightly lower [53]. This tendency can be explained by the fact that the organic part is less cross-linked than the inorganic one. Hardness and Young’s modulus are thus logically lower when the organic content increases. It can be noticed that no other references were found reporting the organic/inorganic effect of an organoalkosilane (without additional organic resin) on mechanical properties on PC. Nevertheless, this study was realized on other substrates, and the same tendency is observed; the organic content tends to decrease H, E [54, 55] and also the scratch resistance [56]. Second, many studies are dedicated to the addition of independent organic components to basic organosilanes to perform a double O/I reticulation within the coating. Epoxide, polyester or polyurethane resins are some example of polymers frequently used. The interest is that

J Sol-Gel Sci Technol

those resins are applied in a wide range of application (coatings, adhesives, resin matrices, etc), they are adherent to many substrates, and they are also temperature and solvent resistant. As logically expected, reported studies [4, 22, 23, 57, 58] show a decrease in hardness with an increasing (resin)/(inorganic part) ratio, since the organic part bring ductility and flexibility and inorganic network cross-linking becomes less effective. Altintas et al. [23] prepared a series of UV-curable I/O hybrid based on TEOS and MEMO as silica precursors and maleimide-modified epoxy resin and urethane acrylate oligomer as organic content. Authors discuss abrasion resistance which is only related to hardness, parameter described as the most influencing factor. Indeed, with the increasing inorganic content of the material, hardness and abrasion resistance are improved. The group of Gungor reports the same conclusion using close systems [4, 22]. Another example, reported by Wouters et al. [24], concerns a system based on acrylate, polyurethane resin, MEMO and TEOS. First, the authors studied the silica content effect. It is mentioned that the Young modulus and hardness increase from 1.8 to 3 GPa and 80 to 200 MPa, respectively, when silica content increases from 0 to 18 wt%. Moreover, incorporation of TEOS and MEMO greatly improves abrasion resistance which increases with an increasing amount of inorganic material. The high abrasion resistance is attributed to the Si–O–Si backbone of the inorganic network. Still considering the abrasion resistance, Kahraman et al. [58] and Gilberts et al. [59] also show that a high condensation rate permits to densify the inorganic network and consequently to improve the abrasion resistance. It can be added that a high degree of condensation can be achieved by modulating the catalyst and the precursors; this point will be discussed in more detail in the ‘‘hybrid coating’’ part. Finally, Shin et al. [60] mention the importance to control the balance between organic and inorganic networks to modulate hardness and confirm that hard coatings are more abrasion resistant. In the same way, Young’s modulus logically decreases with the organic content increasing [4, 23]. A better compatibility between organic and inorganic network tends to increase the Young modulus [4]. To conclude, it is primordial to control the O/I network formation to modulate mechanical properties of the film.

different colloidal silica amount display slight different behavior under scratch solicitation with a Rockwell indenter (200 lm radius): the first crack is observed at higher force (1.4 N) when colloidal silica amount is lower (10 wt%) compared to 1.1 N with the higher colloidal silica amount (50 wt%) (Fig. 4). This evolution can be logically explained by the fact that samples with a higher proportion of hybrid precursors are less brittle, more flexible and thus more elastic [13]. It is also specified by Lionti et al. that scratch resistance does not depend on the origin of the silica (colloidal silica or TEOS) since the groove width remains similar in all the samples. Still concerning the LC1 parameter, Sowntharya et al. who work on a mixed Si/Zr-based system also notice that this value does not change (1 N) for colloidal silica weight ratio varying between 0 and 25 wt%. However, when LC2 (delamination) is studied, conclusions are quite different since it increases with the colloidal silica amount until 10 wt% (Fig. 5). It can be mentioned that after this weight ratio of silica nanoparticles, cracks appear within the coating and LC2 decreases as well as the abrasion resistance. It is explained by the formation of silica clusters within the network which induces a poor compatibility with the matrix [62], and the coating becomes more brittle [13]. Considering now the pencil hardness, Wu et al. [9] and Chen el al. [61] show the same tendency; e.g., pencil hardness grade increases significantly by increasing the silica content into GPTMS/TEOS-based sol. Relationship between colloidal silica amount and coating mechanical properties is confirmed by nanoindention results reported by Lionti et al. [50], Chen et al. [61] and Wu et al. [9] who prove that Young’s modulus and hardness values increase with the colloidal silica content [9, 50] (Fig. 5). Referring to the work of Leyland [45], we propose, as shown in Fig. 6, to calculate and report the H/E ratio which qualifies the ‘‘elastic strain to failure.’’ This

3.2 Filler additivation Another way to tailor mechanical properties of coatings is to add fillers into the sol. It is well known that filler additivation within the network increases both coating hardness and mechanical resistance. Colloidal silica is the most common used additive in sol–gel process [9, 13, 50, 61, 62]. First, Lionti et al. show that samples prepared with

Fig. 5 LC1 and LC2 results with different colloidal silica amount [13, 62]

123

J Sol-Gel Sci Technol

value slightly decreases when colloidal silica amount increases for works reported by Wu et al. [9] and Chen et al. [61]. Concerning the work of Lionti et al. [13], the H/ E ratio behavior is quite equivalent still with a low value, which remains stable with the colloidal silica ratio increasing (Fig. 6). Both results confirm the poor wear resistance of the coatings. Literature also reports another example of fillers which concerns the addition of colloidal TiO2 into silica-based sol. First example, Dinelli et al. [63] show that, incorporating 5 wt% nano-TiO2 into a TEOS-based sol, the LC1 value measured thanks to scratch test with a 800-lm Rockwell tip radius reaches 13.7 ± 0.7 N compared to 7.9 ± 0.6 N for pure silica-based coating. In a second example, Hwang et al. [64] report for silica-based coating that pencil hardness can reach 6H with 5 wt% of nanoTiO2. It can also be noticed that abrasion resistance is enhanced by increasing nano-TiO2 content. 3.3 Mixed SiO2–(Ti/Zr/B/Zn)Ox coating Mixed oxide system is performed to bring new functionalities to the final recovered device. Even if mechanical property enhancement is not directly targeted here, there are still important to master. Indeed, this part will discuss about the effect of other metallic alkoxide addition into the silica-based sol on the coating mechanical properties. First of all, few words about sol synthesis have to be considered, kinetic reactions being closely dependent on the metallic precursor used. Actually, zirconium- and titanium-based alkoxides react faster than silicon-based ones, and in order to control the hydrolysis/condensation reactivity, chelation of the metallic center is often needed with acetylacetone (Acac) or methacrylic acid (MAA) [62, 63, 65]. Consequently, a good mastery of both kinetic reactions of silanes

Fig. 6 Nanoindentation results with different colloidal silica amount [9, 50]

123

and metallic precursors is necessary to avoid any sol destabilization. The nature of hybrid mixed oxide will act on coating mechanical properties. For example, incorporation of boric acid into a silica-based sol creates a Si–O–B network [57]. Pendulum hardness and abrasion resistance are better when B/Si ratio is higher, suggesting an improvement of the cross-linking density when boron is added. On the contrary, Bail et al. [66] report that addition of zirconium-based precursor into a GPTMS-based sol tends to decrease hardness and Young’s modulus values. An explanation is linked to the fact that zirconium-based precursor catalyzes the ring epoxy opening [20, 21, 67] and consequently leads to the formation of an organic network (polyethylene chain) via the latter opened ring epoxy [68]. In spite of that, scratch resistance is revealed to be much better. Actually, LC2 is increased by a factor 2 only by incorporating zirconium propoxide (ZTP) into a silica sol [66]. Dinelli et al., who investigate different titania/silica ratios, observe the same scratch enhancement behavior when tetraethyl orthotitanate (TEOT) is added into a TEOS-based sol. Actually, LC1 value increases with the addition of TEOT until 5 wt%. From 5 % to 100 wt% of TEOT, this beneficial effect is no longer observed since the measured LC1 value decreased to reach a value worse than one reported for uncoated PC [63]. For the latter coating (100 wt% TEOT), SEM images reveal surface defects and cracks; the coating seems to be more brittle. A possible explanation given by the authors is linked to the fast hydrolysis kinetic of TEOT.

4 Process parameters influence 4.1 Thickness Thickness is a crucial parameter when a coating is studied. Many studies in sol–gel field report a close relationship between coating thickness and its mechanical properties. However, conclusions are not so general but really system dependent. First, Yavas et al. find that microhardness value (measured in Vickers HV1) tends to be modified linearly with the coating thickness. Higher hardness values are measured for thicker coatings, from 30 HV1 to 250 HV1 for a thickness around 2.2–8.0 lm, respectively. In this system, it is suggested that fast solvent evaporation occurring for thin coating leads to the formation of pores (detected by SEM) [7]. Indeed, in this specific case, hardness changes are more linked to the coating density than to its thickness. On the contrary, Wu et al. confirm, using nanoindentation, that coating intrinsic hardness is not linked to thickness but only to bulk properties. This theory is confirmed by many other groups

J Sol-Gel Sci Technol

on different substrates [69–71]. However, when pencil hardness is considered, higher values are measured with thick coating inducing a better scratch resistance [9, 18, 61]. Fabbri et al. [72] add that an increase in LC1 and LC2 is observed when thickness increases from 0.085 to 0.125 lm and tends to decrease for thicker coatings. Actually, it can be assumed that above a critical thickness, residual stress in the coating appears leading to the formation and propagation of cracks into the bulk to the interface. Thus, an optimum thickness is needed to reach the best scratch resistance. Finally, abrasion resistance tends to be better with thicker coating as Sowntharya et al. [73] who work on hybrid Si/Ti sol suggest by comparing thick GPTMS-based (GB) coating with thin MPTMS-based (MB) coating. The problem in this specific case is the difficulty to conclude because of the difference of hardness between both coatings, around 0.1 and 0.5 GPa for GB and MB coatings, respectively [18]. Thus, the abrasion resistance could be attributed to thickness as it is mentioned as well as surface or interface properties. Actually, this is not in accordance with Gilbert et al. [59] who find no correlations between abrasion resistance and thickness (range from 10 to 20 lm).

4.2 Aging time It is well known that sol viscosity can be modulated varying either sol aging or hydrolysis ratio. Indeed, by increasing sol viscosity, coating thickness increases in the same way, all implementation parameters being equal. Li et al. [74] notice that by increasing the OH/SiOR ratio from 1.1 to 4.4, transmittance after 500 cycles of abrasion (using a pair of CS10 wheels, 0.5 kg load each) increases from 86 to 98 % indicating a clear abrasion resistance enhancement. Same tendency is observed with hydrolysis time increasing (from 3.5 to 7.5 h). Authors do not claim any explanation for those results. However, one hypothesis can be given since in both cases either OH/SiOR ratio or hydrolysis time increases, condensation rate also increases ensuring abrasion resistance enhancement as mentioned above. In another context, Lionti et al. [13] demonstrate a close relationship between sol aging and scratch resistance. Actually, LC1 and LC2 observed after 1.5 day old are 1.2 and 1.6 N, respectively, whereas it is only 0.4 and 0.8 N, respectively, for a 77-day-old sol. Bulk properties are responsible for LC1 and a possible explanation given by the authors for LC2 is the decrease in adherence with increasing aging time, condensation reactions being advanced for ‘‘old’’ sol and consequently less bonds are available to react with the substrate. Adhesion test, crosscut test or DCB could reinforce this supposition.

4.3 Curing process Literature reports three methods for sol–gel coating curing when deposited on polycarbonate; resistive oven, ultraviolet (UV) and microwave (MW) curing. The most traditional remains the oven curing with a treatment typically ranged from 80 to 135 °C [7, 13, 62, 72], necessary inferior to 140 °C to preserve polycarbonate. Since oven curing tends to accelerate sol–gel reactions by a faster solvent evaporation, it is an important parameter to control. It influences hardness, which is higher with a long treatment [74]. Fabbri et al. also confirm that thermal curing is effective in promoting the cross-linking density of the hybrid system. Thus, crosslinking density being linked to surface hardness and the latter correlated with the abrasion resistance, increasing the curing time leads to the enhancement of all those parameters [72]. It has to be mentioned that oven curing time versus mechanical properties studies are poorly informed on PC, but the same tendency is observed on other substrate like glasstreated at low temperature (80–140 °C) [72]. UV curing is an interesting alternative when polymer substrate is used since it does not need high temperature to generate an organic network. UV curing is commonly used when addition of a UV-curable organoalkoxysilane (MEMO) is used alone or mixed with a UV-curable organic component (acrylate resin for example) to perform a double O/I reticulation within the final coating (Fig. 7) [23–25, 58, 59, 75]. UV curing is often coupled with the traditional oven curing to optimize reticulation degree and then mechanical properties [6, 16, 17, 22, 58, 62, 76]. Even if numerous works report the use of UV-curing for sol–gel UV-curable system on polycarbonate, researchers use this process to densify the network but do not study relationship between UV-curing conditions and mechanical properties of the final coating. Finally, MW-assisted curing is also investigated opening new possibilities. This technique, time and energy saving, is particularly interesting to use with PC, which is transparent to microwave. Another interesting point comes from the MW selectivity to heat only dielectric material, such as silica. Fabbri et al. [2] prove that 10 s is needed with MW against 40 min at 80 °C to reach the same cross-linking rate. Moreover, Dinelli et al. [63] prove that only 120 s is needed to obtain resistant and effective surface coating, LC1 can reach 14 N with a 800 lm tip radius without any post-curing step. 4.4 Substrate/coating interface Finally, another key parameter influencing mechanical behavior of any coating concerns the interface between substrate and coating and especially its adherence quality. In this context, a surface treatment (chemical, plasma, adherence promoter, etc) of the PC is commonly reported in order

123

J Sol-Gel Sci Technol O O

O Si

UV

O

O

O Si

O n

O

O

O

O

O

Fig. 7 Hydrolysis/condensation and UV reactions on MEMO precursor [62, 77]

to enhance it [4, 74]. Moreover, if this implementation is usual, studies reporting surface treatment effect on mechanical properties of the final coating are infrequent. First example concerns works of Tsukakoshi et al., who show an increased abrasion resistance by using an intermediate layer before coating deposition to promote the adherence (confirmed by tape test). Authors report a transmittance, after 140 turns of abrasion tester with steel wool under loading 1 kg, of approximatively 70 %. It is also specified that the promoting agent used which is 3-aminopropyltrimethoxysilane (3-APMS) does not play a role in the finale Martens hardness of the coating. Thus, this enhancement is only due to the better adherence [76]. By coupling those results with work of Lionti et al., it is possible to conclude that adherence is a key parameter to design an abrasion- and scratch-resistant coating. Actually in their former study, Lionti et al. [13] point out a real benefit of an atmospheric plasma (N2) treatment of the PC to enhance the LC2 scratch behavior of GPTES/TEOS-based coating. The authors suggest that covalent bonds are created between epoxy group coming from GPTES and NH2 function formed by plasma on the polycarbonate. This adherence enhancement is confirmed by a latter work, in which Lionti et al. [50] show an increasing by ten of the Gc value (measured by DCB) from a coating deposited on a non-treated PC with a coating deposited on N2 plasma-treated PC. Another way to modulate adherence of the coating to the PC is to add into the silica-based sol a coupling agent able to covalently link the chemical network to the substrate. Moreover, it can be mentioned that adherence influences pencil hardness, since the latter is closely linked to abrasion. Actually, Mizuta et al. [53] report an increased coating adhesion when synthesizing a PhTES/TEOS coating. Interestingly, pencil hardness increases from B to 2B when adhesion (measured by crosscut test) increases from 0 to 100 %. Authors suggest that phenyl group of silane interacts with benzyl group of PC giving p–p interactions at the interface.

5 Conclusions A perfect mastery of the sol chemistry as well as all parameters implemented to prepare a silica-based coating on polycarbonate is essential to control all mechanical

123

properties of the final device. Coatings obtained thanks to the sol–gel process offer multiple possibilities in terms of mechanical performances. The chemistry effect on the final coating properties was clearly demonstrated. Actually, it was shown that hardness and Young’s modulus depend on the organic/inorganic ratio which are lowered when the organic part is higher. Moreover, filler additivation like nano-SiO2 or TiO2 improves both scratch and abrasion resistance when added to an optimum amount. The role of mixed coatings was also clearly identified; for example, Si/ B coating tends to be resistant face to abrasion. Actually, an improvement of the cross-linking density is observed when boron is added. On the contrary, Zr addition into a GPTMS-based sol is reported to decrease hardness and Young’s modulus values. The metal associated with the silica network is thus chosen according to the aimed property. When physical parameters are studied, it is more difficult to draw a tendency since the influence of thickness, sol aging or curing process are really system dependent. To conclude, it is not always easy to draw general conclusions because of the interdependence of all chemical and physical parameters. Nevertheless, abrasion, wear or scratch resistance are controlled by both bulk properties and interface themselves linked to sol chemistry, thickness and adherence. This review proves the importance of an interdisciplinary research between organic/inorganic chemistry and tribology. Moreover, the sol–gel process opens a lot of possibilities to answer to divert technical specifications. Acknowledgments This work was supported by the Programme Avenir Lyon Saint-Etienne (ANR-11-IDEX-0007) of Universite´ de Lyon, within the program ‘‘Investissements d’Avenir’’ operated by the French National Research Agency (ANR) and by the iMUST program.

References 1. Diepens M, Gijsman P (2007) Photodegradation of bisphenol A polycarbonate. Polym Degrad Stab 92:397–406 2. Fabbri P, Leonelli C, Messori M et al (2008) Improvement of the surface properties of polycarbonate by organic–inorganic hybrid coatings. J Appl Polym Sci 108:1426–1436 3. Xia H, Suo ZY, Qiang GJ, Chang CJ (2008) Synthesis, characterization, and degradation of a novel L-tyrosine-derived

J Sol-Gel Sci Technol

4.

5.

6.

7.

8. 9.

10.

11. 12.

13. 14.

15.

16.

17.

18.

19.

20. 21.

22.

23.

polycarbonate for potential biomaterial applications. J Appl Polym Sci 110:2168–2178 Karatas S, Hosgor Z, Kayaman-Apohan N, Gungor A (2009) Preparation and characterization of phosphine oxide containing organosilica hybrid coatings by photopolymerization and sol–gel process. Prog Org Coat 65:49–55 Zhao Y, Ding P, Ba C et al (2014) Preparation of TiO2 coated silicate micro-spheres for enhancing the light diffusion property of polycarbonate composites. Displays 35:220–226. doi:10.1016/ j.displa.2014.06.001 Fateh R, Dillert R, Bahnemann D (2014) Self-Cleaning properties, mechanical stability, and adhesion strength of transparent photocatalytic TiO2–ZnO coatings on polycarbonate. ACS Appl Mater Interfaces 6:2269–2277 Yavas H, Selcuk CDO, Ozhan AES, Durucan C (2014) A parametric study on processing of scratch resistant hybrid sol–gel silica coatings on polycarbonate. Thin Solid Films 556:112–119 Chen DG (2001) Anti-reflection (AR) coatings made by sol–gel processes: a review. Sol Energy Mater Sol Cells 68:313–336 Wu LYL, Chwa E, Chen Z, Zeng XT (2008) A study towards improving mechanical properties of sol–gel coatings for polycarbonate. Thin Solid Films 516:1056–1062 Yahyaei H, Mohseni M (2013) Use of nanoindentation and nanoscratch experiments to reveal the mechanical behavior of sol–gel prepared nanocomposite films on polycarbonate. Tribol Int 57:147–155 Sanchez C, Livage J (1990) Sol–gel chemistry from metal alkoxide precursors. New J Chem 14:513–521 Sanchez C, Belleville P, Popall M, Nicole L (2011) Applications of advanced hybrid organic-inorganic nanomaterials: from laboratory to market. Chem Soc Rev 40:696–753 Lionti K, Toury B, Boissiere C et al (2013) Hybrid silica coatings on polycarbonate: enhanced properties. J Sol-Gel Sci Technol 65:52–60 Chang C-C, Oyang T-Y, Hwang F-H et al (2012) Preparation of polymer/silica hybrid hard coatings with enhanced hydrophobicity on plastic substrates. J Non-Cryst Solids 358:72–76 Zhang S, Chen Z, Guo M et al (2014) Waterborne UV-curable polycarbonate polyurethane nanocomposites based on polydimethylsiloxane and colloidal silica with enhanced mechanical and surface properties. RSC Adv 4:30938–30947 Han Y-H, Taylor A, Knowles KM (2009) Scratch resistance and adherence of novel organic–inorganic hybrid coatings on metallic and non-metallic substrates. Surf Coat Technol 203:2871–2877 Gururaj T, Subasri R, Raju KRCS, Padmanabham G (2011) Effect of plasma pretreatment on adhesion and mechanical properties of UV-curable coatings on plastics. Appl Surf Sci 257:4360–4364 Sowntharya L, Lavanya S, Chandra GR et al (2012) Investigations on the mechanical properties of hybrid nanocomposite hard coatings on polycarbonate. Ceram Int 38:4221–4228 Li CH, Wilkes GL (1998) The mechanism for 3-aminopropyltriethoxysilane to strengthen the interface of polycarbonate substrates with hybrid organic–inorganic sol–gel coatings. J Inorg Organomet Polym 8:33–45 Philipp G, Schmidt H (1986) The reactivity of TiO2 and ZrO2 in organically modified silicates. J Non-Cryst Solids 82:31–36 Kiruthika P, Subasri R, Jyothirmayi A et al (2010) Effect of plasma surface treatment on mechanical and corrosion protection properties of UV-curable sol–gel based GPTS-ZrO2 coatings on mild steel. Surf Coat Technol 204:1270–1276 Basturk E, Inan T, Gungor A (2013) Flame retardant UV-curable acrylated epoxidized soybean oil based organic–inorganic hybrid coating. Prog Org Coat 76:985–992 Altintas Z, Karatas S, Kayaman-Apohan N, Gungor A (2011) The maleimide modified epoxy resins for the preparation of UVcurable hybrid coatings. Polym Adv Technol 22:270–278

24. Wouters MEL, Wolfs DP, van der Linde MC et al (2004) Transparent UV curable antistatic hybrid coatings on polycarbonate prepared by the sol–gel method. Prog Org Coat 51:312–320 25. Belon C, Chemtob A, Croutxe-Barghorn C et al (2011) A simple method for the reinforcement of UV-cured coatings via sol–gel photopolymerization. Macromol Mater Eng 296:506–516 26. Buckle H (1971) Use the hardness test to determine other material properties. In: Westbrook JH, Conrad ASM (eds) Science of hardness testing and its research applications, vol 27. Metals Park, Ohio, pp 453–491 27. Iost A, Bigot R (1996) Hardness of coatings. Surf Coat Technol 80:117–120 28. Landel RF, Nielsen LE (1993) Mechanical properties of polymers and composites, 2nd edn. CRC Press, Boca Raton 29. Oliver WC, Pharr GM (1992) An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. J Mater Res 7:1564–1583 30. Doerner MF, Nix WD (1986) A method for interpreting the data from depth-sensing indentation instruments. J Mater Res 1:601–609 31. VanLandingham MR, Villarrubia JS, Guthrie WF, Meyers GF (2001) Nanoindentation of polymers: an overview. Macromol Symp 167:15–43 32. Martens A (1898) Handbuch der materialienkunde fu¨r den maschinenbau: Materialpru¨fungswesen, probirmaschinen und messinstrumente. Springer, Berlin 33. Pethicai JB, Hutchings R, Oliver WC (1983) Hardness measurement at penetration depths as small as 20 nm. Philos Mag A 48:593–606 34. Loubet J, Georges J, Marchesini O, Meille G (1984) Vickers indentation curves of magnesium-oxide (Mgo). J Tribol Trans ASME 106:43–48 35. Bec S, Tonck A, Loubet JL (2006) A simple guide to determine elastic properties of films on substrate from nanoindentation experiments. Philos Mag 86:5347–5358 36. Han SM, Saha R, Nix WD (2006) Determining hardness of thin films in elastically mismatched film-on-substrate systems using nanoindentation. Acta Mater 54:1571–1581 37. Li XD, Bhushan B (2002) A review of nanoindentation continuous stiffness measurement technique and its applications. Mater Charact 48:11–36 38. Bull SJ (1997) Failure mode maps in the thin film scratch adhesion test. Tribol Int 30:491–498 39. Hedenqvist P, Hogmark S (1997) Experiences from scratch testing of tribological PVD coatings. Tribol Int 30:507–516 40. Benayoun S, Fouilland-Paille L, Hantzpergue JJ (1999) Microscratch test studies of thin silica films on stainless steel substrates. Thin Solid Films 352:156–166 41. Archard J (1953) Contact and rubbing of flat surfaces. J Appl Phys 24:981–988 42. Magnee A (1993) Modelization of damage by abrasion. Wear 162:848–855 43. Bull SJ (1999) Can scratch testing be used as a model for the abrasive wear of hard coatings? Wear 233:412–423 44. Kato K (1995) Microwear mechanisms of coatings. Surf Coat Technol 76–77:469–474 45. Leyland A, Matthews A (2000) On the significance of the H/E ratio in wear control: a nanocomposite coating approach to optimised tribological behaviour. Wear 246:1–11 46. Musil J, Kunc F, Zeman H, Polakova H (2002) Relationships between hardness, Young’s modulus and elastic recovery in hard nanocomposite coatings. Surf Coat Technol 154:304–313 47. Tsui TY, Pharr GM, Oliver WC et al (1995) Nanoindentation and nanoscratching of hard carbon coatings for magnetic disks. In: Symposium I—Mechanical Behavior of Diamond and Forms of Carbon

123

J Sol-Gel Sci Technol 48. Pharr GM (1998) Measurement of mechanical properties by ultralow load indentation. Mater Sci Eng, A 253:151–159 49. Tsui TY, Pharr GM (1999) Substrate effects on nanoindentation mechanical property measurement of soft films on hard substrates. J Mater Res 14:292–301 50. Lionti K, Cui L, Volksen W et al (2013) Independent control of adhesive and bulk properties of hybrid silica coatings on polycarbonate. ACS Appl Mater Interfaces 5:11276–11280 51. Cui L, Ranade AN, Matos MA et al (2012) Atmospheric plasma deposited dense silica coatings on plastics. ACS Appl Mater Interfaces 4:6587–6598 52. Le Houe´rou V, Gauthier C, Schirrer R (2010) Mechanical analysis of the blistering of a thin film deposited on a glassy polymer. Tribol Int 43:129–135 53. Mizuta Y, Daiko Y, Mineshige A et al (2011) Phase-separation and distribution of phenyl groups for PhTES-TEOS coatings prepared on polycarbonate substrate. J Sol-Gel Sci Technol 58:80–84 54. Lee CH, Lee SG, Lee JD (2012) Preparation and characterization of photochromic organic–inorganic hybrid coating using 1,2bis(2,4-dimethyl-5-phenyl-3-thienyl)3,3,4,4,5,5-hexafluoro-1-cyclope ntene. Polym Korea 36:16–21 55. Trabelsi O, Tighzert L, Jbara O, Hadjadj A (2011) Synthesis via sol–gel process and characterization of novel organic–inorganic coatings. J Non-Cryst Solids 357:3910–3916 56. Grundwu¨rmer M, Nuyken O, Meyer M et al (2007) Sol–gel derived erosion protection coatings against damage caused by liquid impact. Wear 263:318–329. doi:10.1016/j.wear.2006.12. 039 57. Mulazim Y, Kahraman MV, Apohan NK et al (2011) Preparation and characterization of UV-curable, boron-containing, transparent hybrid coatings. J Appl Polym Sci 120:2112–2121 58. Kahraman MV, Bayramoglu G, Boztoprak Y et al (2009) Synthesis of fluorinated/methacrylated epoxy based oligomers and investigation of its performance in the UV curable hybrid coatings. Prog Org Coat 66:52–58 59. Gilberts J, Tinnemans AHA, Hogerheide MP, Koster TPM (1998) UV curable hard transparent hybrid coating materials on polycarbonate prepared by the sol–gel method. J Sol-Gel Sci Technol 11:153–159 60. Shin YJ, Yang DH, Oh MH et al (2009) Hard coatings on polycarbonate plate by sol–gel reactions of melamine derivative, poly(vinyl alcohol), and silicates. J Ind Eng Chem 15:238–242 61. Chen Z, Wu LYL, Chwa E, Tham O (2008) Scratch resistance of brittle thin films on compliant substrates. Mater Sci Eng, A 493:292–298 62. Sowntharya L, Gundakaram RC, Raju K, Subasri R (2013) Effect of addition of surface modified nanosilica into silica-zirconia hybrid sol–gel matrix. Ceram Int 39:4245–4252

123

63. Dinelli M, Fabbri E, Bondioli F (2011) TiO2–SiO2 hard coating on polycarbonate substrate by microwave assisted sol–gel technique. J Sol-Gel Sci Technol 58:463–469 64. Hwang DK, Moon JH, Shul YG et al (2003) Scratch resistant and transparent UV-protective coating on polycarbonate. J Sol-Gel Sci Technol 26:783–787 65. Del Angel-Lopez D, Dominguez-Crespo MA, Torres-Huerta AM et al (2013) Analysis of degradation process during the incorporation of ZrO2: SiO2 ceramic nanostructures into polyurethane coatings for the corrosion protection of carbon steel. J Mater Sci 48:1067–1084 66. Le Bail N, Lionti K, Benayoun S et al (2015) Scratch resistant sol–gel coatings on pristine polycarbonate. New J Chem (accepted manuscript) 67. Innocenzi P, Esposto M, Maddalena A (2001) Mechanical properties of 3-glycidoxypropyltrimethoxysilane based hybrid organic-inorganic materials. J Sol-Gel Sci Technol 20:293–301 68. Blanc D, Last A, Franc J et al (2006) Hard UV-curable organomineral coatings for optical applications. Thin Solid Films 515:942–946 69. Kessman AJ, Huckaby DKP, Snyder CR et al (2009) Tribology of water and oil repellent sol–gel coatings for optical applications. Wear 267:614–618 70. Robertson MA, Rudkin RA, Parsonage D, Atkinson A (2003) Mechanical and thermal properties of organic/inorganic hybrid coatings. J Sol-Gel Sci Technol 26:291–295 71. Hu L, Zhang X, Sun Y, Williams RJJ (2005) Hardness and elastic modulus profiles of hybrid coatings. J Sol-Gel Sci Technol 34:41–46 72. Fabbri P, Messori M, Toselli M et al (2008) Enhancing the scratch resistance of polycarbonate with poly(ethylene oxide)silica hybrid coatings. Adv Polym Technol 27:117–126 73. Raju KRCS, Sowntharya L, Lavanya S, Subasri R (2012) Effect of plasma pretreatment on adhesion and mechanical properties of sol–gel nanocomposite coatings on polycarbonate. Compos Interfaces 19:259–270 74. Li CH, Jordens K, Wilkes GL (2000) Abrasion-resistant coatings for plastic and soft metallic substrates by sol–gel reactions of a triethoxysilylated diethylenetriamine and tetramethoxysilane. Wear 242:152–159 75. Oktay B, Kayaman-Apohan N (2013) Maleimide containing UVcured hybrid coatings. Adv Polym Technol 32:21341 76. Tsukakoshi S, Itatani K, Koda S (2011) Properties of silicious film on polycarbonate substrate prepared by vacuum ultraviolet irradiation: effect of intermediate silane layer. J Eur Ceram Soc 31:2489–2496 77. Han Y-H, Taylor A, Mantle MD, Knowles KM (2007) UV curing of organic–inorganic hybrid coating materials. J Sol-Gel Sci Technol 43:111–123