- Email: [email protected]
Effect of thermal history on scratch behavior of multi-phase styrenic-based copolymers
Tribology International 99 (2016) 248–257
Contents lists available at ScienceDirect
Tribology International journal homepage: www.elsevier.com/locate/triboint
Effect of thermal history on scratch behavior of multi-phase styrenic-based copolymers Mohammad Motaher Hossain a, Eike Jahnke b, Philipp Boeckmann c, Svetlana Guriyanova c, Rolf Minkwitz c, Hung-Jue Sue a,n a
Polymer Technology Center Department of Materials Science and Engineering, Texas A&M University, College Station, TX 77843, USA Styrolution Group GmbH, 60325 Frankfurt, Germany c BASF SE, Advanced Materials & Systems Research, 67056 Ludwigshafen, Germany b
art ic l e i nf o
a b s t r a c t
Article history: Received 13 November 2015 Received in revised form 23 March 2016 Accepted 24 March 2016 Available online 31 March 2016
Effect of thermal history on scratch performance of acrylonitrile styrene acrylate (ASA) and acrylonitrile butadiene styrene (ABS) model systems was studied. ASAs with rubber particle sizes of 100 nm and 1 mm, respectively, and ABS with rubber particle size of 100 nm were investigated. Linearly increasing normal load scratch tests were performed according to the ASTM-D7027/ISO-19252 standard. An improvement in scratch resistance is observed in ASAs when annealed at 140 °C whereas a noticeable drop in scratch resistance is found in ABS under the same condition. Various microscopic analyses were conducted to determine the corresponding deformation mechanisms for the observed dependence of scratch performance on annealing. Implication of the present findings for design of scratch resistant rubber toughened polymers is discussed. & 2016 Elsevier Ltd. All rights reserved.
Keywords: Rubber modification Heat treatment Scratch resistance Damage mechanisms
1. Introduction Scratch behavior of multi-phase polymeric systems is gaining significant interests due to its increasingly wide usage in engineering, and structural and durable goods applications. However, in-depth understanding of scratch-induced surface deformation in multi-phase polymeric systems is quite challenging since the presence of micro- and nano-phase domains near the surface can greatly influence the scratch deformation and damage [1]. It is not straightforward to relate the surface deformation and damage phenomena to bulk mechanical properties in multi-phase polymeric systems. For instance, rubber addition to the matrix improves toughness and ductility but simultaneously reduces modulus and strength [2,3], which complicates the fundamental understanding of their influences on scratch behavior. For homogeneous single-phase polymers, researchers have shown some success in correlating material and surface properties to the evolution of different scratch-induced surface damage features [4–24]. Using the ASTM D7027/ISO 19252 scratch test standard [25], it has been shown that an increase in tensile strength delays the onset of microcrack formation during scratching in styrene-acrylonitrile (SAN) random copolymers [5]. Bucaille et al. n
Corresponding author. Tel.: þ 1 979 845 5024; fax: þ 1 979 845 3081. E-mail address: [email protected] (H.-J. Sue).
http://dx.doi.org/10.1016/j.triboint.2016.03.026 0301-679X/& 2016 Elsevier Ltd. All rights reserved.
[7] showed that a more pronounced strain hardening can improve scratch resistance, defined as the resistance to surface deformation and damage due to sliding indentation of a rigid asperity, since it restricts large-scale plastic deformation. Using the finite element method (FEM) parametric study and experimental observation, it has been shown that, in addition to the coefficient of friction, yield stress, strain at stress recovery and strain hardening slope in compression are the most important material parameters that determine the development of scratch groove during scratching [9–11]. Tensile behavior has little influence on scratch groove formation but correlates well with microcrack formation during scratching [9–11]. In a recent study [26,27], it has been shown that by knowing the material behavior and surface characteristics, the scratch behavior of polymers can be quantitatively predicted with reasonable accuracy. For multi-phase polymeric systems, understanding and analysis of the scratch behavior is rather complex. In-depth understanding of the localized deformation and damage due to interaction between the micro-/nano-sized particles and matrix is needed. Introduction of a rubber phase in a polymer matrix has shown to significantly alter the stress state near the rubber particles [28–31], which could ultimately change the deformation mechanisms. Browning et al. [32] showed that the crystallinity of the ethylene segment and the internal morphology of the ethylene-propylene rubber (EPR) phase significantly affect the
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
scratch behavior of soft thermoplastic olefins (TPOs). In another study [33], it has been shown that the degree of surface crystallinity, which can be modified by varying the processing conditions, can significantly influence the scratch behavior of polypropylene (PP). It was further shown that [34] improvements in different bulk mechanical properties by adding synthetic clay nanoplatelets or core-shell rubber (CSR) nanoparticles in epoxy do not necessarily enhance the scratch resistance. Liang et al. [35] showed that the scratch resistance of rubber-modified acrylonitrile styrene acrylate (ASA) system deteriorates with increasing rubber content due to reduction in tensile and compressive properties. In a more recent study [1], Hossain et al. reported a reduction in scratch resistance in ASA containing 100 nm rubber particles when compared to ASA containing 1 mm rubber particles. It was further shown that rubber type has significant influence on scratch behavior as acrylonitrile butadiene styrene (ABS) exhibits a higher scratch resistance when compared to ASA with comparable rubber particle size. Since the bulk mechanical behaviors in tension and compression of the model systems are similar, they were not so useful in explaining the observed difference in scratch resistance in the rubber-modified systems. It was argued [1] that the rubber particle type and size can alter the damping characteristics of the SAN matrix, hence the frictional behavior to cause variation in their scratch behavior. Thermal history has long been shown to significantly affect the physical and mechanical properties of different polymers. Variation in processing condition can induce and alter skin-core characteristics and significantly affect the final morphology, especially on sample surfaces. However, for styrenic-based copolymers, it has been shown that the processing conditions have insignificant effect on mechanical properties [36,37]. In particular, tensile yield stress and Charpy impact strength have been shown to remain the same upon variation in heat treatment procedure. It was also shown that [37] high temperature annealing can improve the scratch resistance without influencing the bulk properties in rubber-modified styrene-acrylonitrile (SAN) systems. The present study focuses on understanding how the thermal history influences the scratch behavior of styrenic-based Table 1 Material information of model systems.
ASA100 ABS100 ASA1000
Rubber type
Structure type
Rubber particle size
ASA ABS ASA
SAN grafted PBA rubber SAN grafted PBD rubber SAN grafted PBA rubber
E100 nm E100 nm E1 mm
249
copolymers. The model systems, acrylonitrile styrene acrylate (ASA) and acrylonitrile butadiene styrene (ABS) are composed of SAN matrix, and acrylic and butadiene rubbers, respectively. Extensive analysis has been performed to understand the observed differences in scratch resistance and evolution of different scratchinduced surface deformation mechanisms. Implications of the present study for designing scratch resistant rubber-modified polymers are discussed.
2. Experimental 2.1. Materials The ASA and ABS copolymers utilized in this study were provided by Styrolution Group GmbH (Frankfurt, Germany). The ASA systems consist of a random copolymer SAN matrix and grafted polybutyl-acrylate (PBA) rubber particles with an average nominal diameter of 100 nm and 1 mm, respectively. Size distribution of the rubber particles in the model systems can be found elsewhere [1]. In the SAN phase, the acrylonitrile content was controlled to be at 35 wt%, and the weight-average molecular weight (Mw) of SAN was chosen to be 104 kg/mol [1] with a polydispersity (PDI) of 3.7. The ABS system consisted of a random copolymer SAN matrix and grafted polybutadiene (PBD) rubber particles with an average nominal diameter of approximately 100 nm. Within the SAN phase the same material as described for ASA was used. The rubber concentration is 30 wt% for all the systems investigated. Test specimens were fabricated by injection molding with a fan gate design to spread and slow the melt as it enters the mold cavity to ensure near-uniform molecular orientation across the width of the plaques. The plaques were 150 mm 150 mm in rectangular shape and 6 mm in thickness. Nomenclature and physical characteristics of each system are listed in Table 1. 2.2. Heat treatment Upon receipt, all the plaques were dried in a vacuum oven for 12 h at 80 °C with a vacuum pressure of 30 mm Hg. The drying process was conducted by sandwiching the plaques in between two smooth glass plates (Surface roughness, Ra ¼17.0 nm, Rq ¼22.3 nm). An external weight leading to stress of about 10 kPa was applied to the sandwiched samples during heating and cooling to prevent the potential for uneven residual stress relief or warping of the plaques. After drying, the high temperature annealing (HTA) process was carried out on one of the two dry reference plaques for each system studied. The HTA process was conducted at 140 °C, which is about
Fig. 1. AFM micrographs of the ASA100 cross-section in the sub-surface, (a) dry reference sample and (b) HTA sample.
250
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
Fig. 2. AFM micrographs of the ABS100 cross-section in the sub-surface, (a) dry reference sample and (b) HTA sample.
Fig. 3. AFM micrographs of the ASA1000 cross-section in the sub-surface, (a) dry reference sample and (b) HTA sample.
30 °C higher than the measured Tg [1], for 2 h under vacuum pressure of 30 mm Hg. Afterwards, the samples were slowly cooled to ambient temperature at 1.8 °C/min. The heat treatment procedure is free from any polymer degradation due to oxidation. All the samples (both Dry reference and HTA) were kept in a desiccator before the scratch tests to minimize moisture adsorption from the atmosphere. 2.3. Scratch test Scratch tests were carried out according to the ASTM D7027 standard [25] by using a progressive normal load range of 1–70 N at a constant scratch speed of 1 mm/s for a length of 100 mm. A stainless steel spherical scratch tip with diameter of 1 mm was used. A minimum of five scratch tests were performed at room temperature on each plaque in the melt flow direction of the sample, in accordance with the ASTM standard to allow for controlled testing and analysis. 2.4. Analysis of scratch-induced damage Analysis of scratch-induced surface damage was carried out 72 h after the completion of scratch tests to allow for sufficient viscoelastic recovery. A Keyences VK9700 violet laser scanning confocal microscope (VLSCM) was used for high-resolution surface topographical analysis. The scratch depth and shoulder height at different locations on the scratch path were measured using the associated VK Analyzer software. The critical normal loads for the onsets of cracking and
plowing were measured, as well. The onset of cracking was determined at the point where periodic repetition of the microcracks in or around the center of the scratch groove started by carefully observing the whole scratch under the microscope. The height profile along the scratch path, generated by the VK Analyzer software, was also analyzed along with the optical images to precisely locate the onset of microcrack formation. Surface roughness of the unscratched sample was also measured to study the effect of thermal history on surface morphology. When reporting the data, error bars in the plots indicate one standard deviation (or 1σ) of the mean value.
2.5. Dynamic mechanical analysis (DMA) Dynamic mechanical analysis in torsional mode was performed on a TA Instruments ARES G2 Rheometer. Dry samples with nominal dimensions of 30 mm 10 mm 3 mm were prepared for the DMA. The samples were cut by an Isomet 1000 Precision Saw and the surfaces of the specimens were polished using 4000 grit silicon carbide polishing paper until a surface roughness of 1 mm or less has been achieved. Strain amplitude of 0.05% was chosen for the analysis. The samples were tested from 135 °C to 180 °C with a temperature increment of 3 °C/min at a fixed frequency of 1 Hz. Storage modulus (G0 ) and loss modulus (G00 ) information was recorded during thetest and corresponding loss 00
tangent value ( tan δ) was calculated tan δ ¼ GG0 .
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
Fig. 4. Plot of a) scratch depth and b) shoulder height, as a function of scratch normal load in ASA100.
251
Fig. 5. Plot of a) scratch depth and b) shoulder height, as a function of scratch normal load in ASA1000.
2.6. Atomic force microscopy (AFM) The characterization of sample morphology on the crosssection in the sub-surface using AFM (Dimension, Veeco) was carried out by BASF SE (Ludwigshafen, Germany). The AFM imaging and analysis were performed in Tapping Mode. The cantilever was excited at its resonance frequency and moved at a defined height over the sample surface. Thus, interaction forces cause a variation in the oscillation amplitude of the lever. Furthermore, a phase shift between the excitation and response oscillation can be induced by different material properties such as stiffness. In addition to the height, this effect provides material contrast information of the sample (Phase images). An Olympus silicon tip with cantilever spring constant of 40 N/m was used. The smooth cross-section surfaces of the samples were prepared using microtome technique at 80 °C. The AFM measurements were carried out at room temperature. 2.7. AFM peak force quantitative nanomechanical property mapping (AFM PF-QNM) Peak force quantitative nanomechanical property mapping (PFQNMs) allows quantitative nanomechanical mapping of material properties, including modulus, adhesion, deformation and dissipation, while simultaneously imaging sample topography at high resolution. The surface morphology of the model systems was characterized in-house using AFM (Bruker Dimension Icon AFM) in peak force Tapping Mode. ScanAsysts, which uses algorithms to automatically and continuously monitor image quality and make
Fig. 6. Critical normal loads for onsets of microcracking in the model systems.
the appropriate parameter adjustments, was used for the imaging purpose. Material properties, such as, modulus and deformation, was mapped on the surface of the model systems. The cantilever spring constant of 40 N/m was used for the AFM measurements, carried out at room temperature.
3. Results and discussion AFM micrographs of the dry reference and annealed samples under cross-sectional view in the sub-surface (ca. 50 mm below the
252
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
Fig. 7. Critical normal loads for onsets of plowing in the model systems.
surface) are shown in Figs. 1–3. In both ASA100 and ABS100, the rubber particle size is approximately 100 nm; whereas in ASA1000, the rubber is 1 mm in size. It can be observed that the thermal history has negligible effect on the shape, size and orientation of the rubber particles in the sub-surface. Figs. 4 and 5 show the plots of evolution of residual scratch depth and shoulder height of the scratch groove as a function of scratch normal load obtained via VLSCM for ASA100 and ASA1000, respectively. The numbers 1 to 5 in the plots denote five individual scratch measurements on the same sample. As discussed, the scratch damages were scanned 72 h after the scratch tests to allow for viscoelastic recovery. All measurements were conducted from low load towards high load up to the occurrence of well-developed cracks as they can obscure the scratch depth measurement. Since ABS100 shows onset of crack formation at very low load, to be discussed later, the scratch depth and shoulder height of the scratch groove was not measured for ABS100. As can be seen in Figs. 4 and 5, the scratch depth and shoulder height are similar despite the difference in thermal history for both ASAs. Thus, the difference in heat treatment has negligible influence on the development of scratch groove along the scratch path. Critical normal loads for onsets of microcracking and plowing along the scratch path for the model systems are shown in Figs. 6 and 7, respectively. All the rubber-modified systems show deterioration of scratch resistance with earlier onsets of damage transitions when compared to the neat system we have studied earlier (For neat system, onset of microcracking: 45 N, onset of plowing: 68 N) [5]. For ASAs, high temperature annealing is observed to improve the scratch resistance by delaying the onset of microcracking. The onset of plowing for ASA1000 is also delayed due to annealing. On the contrary, high temperature annealing has induced an earlier onset of microcrack formation in ABS100 although it slightly delayed the onset of plowing. Thus, the effect of thermal history on scratch resistance is greatly influenced by these two rubber types in the rubber-modified systems. It should be noted that the difference in onset of damage transitions between dry ASA100 and ABS100 model systems is insignificant compared to our previous study [1]. However, the drying procedures in both cases are different. Fig. 8 shows the comparison of representative scratch coefficient of friction (SCOF ¼ FFnt ; where Ft is the tangential force and Fn is the normal load) curves obtained during experiment for the model systems. The SCOF consists of two terms: a conventional frictional term, i.e., coefficient of friction (COF) due to adhesion and an additional term due to material deformation. As can be seen in the figure, dry reference systems show higher SCOF compared to their HTA counterparts for ASAs at low
Fig. 8. Representative scratch coefficient of friction (SCOF) curves for a) ASA100, b) ABS100, and c) ASA1000.
normal load regime. Since the residual scratch depth remains essentially the same for both dry reference and HTA samples in ASAs (Figs. 4 and 5) and it has been shown that the heat treatment process has insignificant influence on bulk mechanical properties [37], the discrepancy in the SCOF can possibly be due to the difference in coefficient of friction due to adhesion (i.e., COF). It has been shown that the onset of microcrack formation depends on COF as it intensifies the tensile stress state on the surface behind the scratch tip [38,39]. Thus, the difference in COF can be considered responsible for earlier onset of microcrack formation in the dry reference samples in ASAs. For
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
253
Fig. 9. Optical image at onset of microcrack formation for a) ASA100, b) ABS100, and c) ASA1000. (Scratch direction: from left to right; arrow indicates the onset point on the track).
Fig. 10. Optical image of crack patterns for a) ASA100, and b) ASA1000. (Scratch direction: from left to right).
254
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
Fig. 11. Transition of crack patterns in ABS100 a) dry reference, and b) HTA. (Scratch direction: from left to right).
Fig. 12. Loss tangent (tan δ) of the model systems as a function of temperature below T g .
ABS100, the SCOF remains essentially the same before the onset load for microcracking in dry reference and HTA samples, unable to explain the observed difference in onset of microcracking. Optical micrographs of the onset of scratch-induced microcrack formation in each model system are shown in Fig. 9. It can be seen that the microcrack patterns differ among the rubber-modified systems. Thermal history has no influence on the crack pattern. For ASAs, the microcrack formation around the onset appears to be caused by tensile fracture as denoted by the opening type of cracks [1,15]. Whereas, for ABS100, the damage pattern in the initial stage can be considered shear dominated mixed-mode in nature as shown by the peeling kind of damage [1]. As scratch normal
Fig. 13. Surface roughness of the model systems.
load increases, the microcracks develop into larger cracks for both ASAs (Fig. 10). However, for ABS100, the damage pattern eventually turns from mixed-mode to tensile dominated fracture (Fig. 11). The damage patterns observed in the dry reference system in this study are consistent with the literature [1]. The reason for the difference in the damage pattern between ASAs and ABS can be due to the difference in miscibility between the rubber particles and the matrix. DMA study on the model systems was conducted to support the conjecture. DMA study of the model systems was carried out to probe their viscoelastic behavior. The tan δ plots in semi-logarithmic scale before the glass transition temperature (Tg) are shown in Fig. 12. Tg, which is defined as the temperature at the maximum peak in tan δ curve, does not change with the presence of rubber particles
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
255
Fig. 14. Modulus mapping on the surface using peak force quantitative nanomechanical property mapping (PF-QNM) for ASA100 a) dry reference, b) HTA, and c) modulus distribution. (Darker regions indicate softer material, brighter regions indicate harder material).
Fig. 15. Modulus mapping on the surface using peak force quantitative nanomechanical property mapping (PF-QNM) for ABS100 a) dry reference, b) HTA, and c) modulus distribution. (Darker regions indicate softer material, brighter regions indicate harder material).
Fig. 16. Modulus mapping on the surface using peak force quantitative nanomechanical property mapping (PF-QNM) for ASA1000 a) dry reference, b) HTA, and c) modulus distribution. (Darker regions indicate softer material, brighter regions indicate harder material).
[1]. The secondary transition denoting the Tg of the rubber particles can be identified in the plots ( 90 °C for ABS and 45 °C for ASAs). Introduction of rubber particles in SAN matrix has altered the damping characteristics of the matrix. ASA100 shows the most pronounced damping behavior change at temperatures above the T g of the butyl-acrylate rubber, i.e., 45 °C. This suggests that the 100 nm size butyl-acrylate rubber influences the damping behavior of the SAN matrix most significantly in the model systems investigated. The T g of the butyl-acrylate rubber also appears to have shifted to a higher temperature in ASA100 compared to
ASA1000, suggesting good interaction between the 100 nm butylacrylate rubber and SAN matrix. On the other hand, for ASA1000, the butyl-acrylate rubber does not seem to exert a similar degree of influence in damping behavior of SAN matrix. Interestingly, the 100 nm size butadiene rubber in ABS100 does not show the same influence on SAN damping behavior as ASA100. This suggests that butadiene rubber is less miscible with SAN [1,40–42]. As a result, the interface between the rubber particles and matrix can be considered sharp and the adhesion between the rubber particles and the matrix is likely to be weak in ABS100 compared to those of the ASAs. These differences can be responsible for the observed
256
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
variation in scratch damage patterns. It should be noted that, no observable degradation is detected due to high temperature annealing process shown in one of the previous studies [37]. Surface roughness may influence the effective contact area during sliding contact, which alters the COF [43,44], and, consequently the scratch behavior [11]. However, if the surface roughness is sufficiently low and the applied stress is high enough to induce large scale plastic deformation through the roughness layer, the roughness effect on COF becomes negligible [43,44]. The similarities in residual scratch depth and shoulder height between dry reference and HTA systems in ASAs (Figs. 4 and 5) show that indeed the change in surface roughness due to annealing process has insignificant influence on scratch-induced deformation and damage. Although the amount of surface roughness change due to annealing has negligible influence on scratch behavior, the topographical change can provide useful insights on molecular relaxation and strain recovery due to annealing. Fig. 13 shows the RMS roughness values of the dry reference and HTA unscratched samples, measured on an area of 525 mm 700 mm (at 20 magnification in VLSCM). The area chosen for surface roughness measurement is comparable with the scratch tip geometry and deformation of interest. As can be seen in the figure, the surface roughness increases considerably due to the annealing process for the ASAs. The increase in surface roughness is more pronounced in ASA100 than in ASA1000. Interestingly, for ABS100, the surface roughness change is minimal. The change in surface roughness can be due to relaxation of the highly stretched molecules near the surface which causes rubber particles to be more exposed to the surface during the annealing process [37], leading to a significant increase in surface roughness. The validity of this conjecture will be explored further in the peak force quantitative nanomechanical modulus mapping on the surface. The difference in the degrees of surface roughness change between the ASAs and ABS is probably due to a more coordinated stretching between SAN molecules and butyl-acrylate rubber particles than between SAN and butadiene rubber particles upon injection molding. Figs. 14–16 show the modulus mapping on the surface obtained from peak force quantitative nanomechanical mapping (PF-QNM) for ASA100, ABS100 and ASA1000, respectively. The dark regions on the images indicate softer or lower modulus material whereas the brighter regions indicate the opposite. The rubber particle size and shape on the surface can be observed from the modulus mapping. The modulus distribution on the surface is calculated by normalizing the matrix modulus obtained from PF-QNM against the neat matrix modulus provided by BASF SE. For ASAs, the modulus distribution is shifted towards lower modulus for rubber particles in HTA sample compared to the dry reference counterpart. This shift is most prominent in ASA1000 system indicating that there might be a coating of matrix material on top of the rubber particles in the dry reference system, which was then partially/fully removed upon high temperature annealing. This phenomenon can be due to the retraction of SAN matrix surrounding the rubber particles during the annealing process. The rubber particles have also recovered some of its injection molding induced orientation. It is noted that the matrix modulus on the surface becomes more distributed in the HTA system compared to that of dry reference sample, suggesting that the annealing process has helped reduce stress concentration at the rubber and matrix interface, which might contribute to a delay in onset of microcracking in the HTA systems in ASAs. For ABS100, however, the mechanism of microcrack formation is completely different. The damage pattern is observed to be mixed-mode fracture which eventually turns into tensile dominated fracture at higher normal loads (Figs. 9–11). As can be seen in Fig. 15, no significant difference can be found between the HTA sample and dry reference system in terms of modulus distribution
on the surface. Since the onset of microcrack formation in ABS100 is due to mixed-mode fracture, the interfacial adhesion between rubber particles and matrix are expected to play a major role. The annealing process can induce a greater relaxation of rubber particles and matrix molecules in ABS100 leading to sharper interface and weaker adhesion compared to the dry reference counterpart. As a result, HTA sample shows an earlier onset of microcracking compared to dry reference system in ABS100 as shown in Fig. 6. The present study suggests that the scratch behavior of rubbermodified systems depends on thermal history. The influence of thermal history on scratch behavior depends on the rubber type. The miscibility of the rubber particles with the matrix and corresponding viscoelastic response can be considered responsible for the observed difference in damage pattern formed. Change in surface morphology due to high temperature annealing can be the reason for observed difference in scratch resistance of the model systems investigated. More in-depth study is required to further investigate the complex stress and strain fields developed around the rubber particles in rubber-modified systems.
4. Conclusions The effect of thermal history was studied on scratch performance of two types of styrenic-based copolymers with variation in rubber type and size. Linearly increasing normal load scratch tests were performed according to the ASTM standard for scratch testing to characterize their scratch performance. Significant change in surface morphology has been observed in the ASAs due to possible retraction of SAN matrix surrounding the rubber particles. On the contrary, ABS shows less retraction around the rubber particles. An improvement in scratch resistance is observed in the ASAs when annealed at high temperature whereas a noticeable drop in scratch resistance is found in ABS upon annealing. Thus, the influence of thermal history on scratch performance of rubber-modified systems has shown to be dependent on the rubber type. Due to multiphase nature of the systems studied, scratch-induced deformation and its relation to the thermal history is rather complicated in the rubber-modified systems, and needs further comprehensive understanding. Additional studies on the systems containing rubber particle sizes between 100–1000 nm and different rubber types could further elucidate the effect of thermal history on scratch behavior of rubber modified systems.
Acknowledgments The authors would like to acknowledge the financial support and valuable insights provided by Styrolution Group GmbH in this research endeavor. Special thanks are also given for providing the model polymer systems.
References [1] Hossain MM, Moghbelli E, Jahnke E, Boeckmann P, Guriyanova S, Sander R, Minkwitz R, Sue H-J. Rubber particle size and type effects on scratch behavior of styrenic-based copolymers. Polymer 2015;63:71–81. [2] Kim H, Keskkula H, Paul DR. Toughening of SAN copolymers by an SAN emulsion grafted rubber. Polymer 1990;31:869–76. [3] Steenbrink AC, Litvinov VM, Gaymans RJ. Toughening of SAN with acrylic coreshell rubber particles: particle size effect or cross-link density? Polymer 1998;39:4817–25. [4] Browning RL, Lim GT, Moyse A, Sun L, Sue H-J. Effects of slip agent and talc surface-treatment on the scratch behavior of thermoplastic olefins. Polym Eng Sci 2006;46:601–8.
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
[5] Browning RL, Sue H-J, Minkwitz R, Charoensirisomboon P. Effects of acrylonitrile content and molecular weight on the scratch behavior of styreneacrylonitrile random copolymers. Polym Eng Sci 2011;51:2282–94. [6] Bucaille JL, Felder E, Hochstetter G. Mechanical analysis of the scratch test on elastic and perfectly plastic materials with the three-dimensional finite element modeling. Wear 2001;249:422–32. [7] Bucaille JL, Gauthier C, Felder E, Schirrer R. The influence of strain hardening of polymers on the piling-up phenomenon in scratch tests: Experiments and numerical modelling. Wear 2006;260:803–14. [8] Chu J, Rumao L, Coleman B. Scratch and mar resistance of filled polypropylene materials. Polym Eng Sci 1998;38:1906–14. [9] Hossain MM, Browning RL, Minkwitz R, Sue H-J. Effect of asymmetric constitutive behavior on scratch-induced deformation of polymers. Tribol Lett 2012;47:113–22. [10] Hossain MM, Jiang H, Sue H-J. Effect of constitutive behavior on scratch visibility resistance of polymers-A finite element method parametric study. Wear 2011;270:751–9. [11] Hossain MM, Minkwitz R, Sue H-J. Minimization of surface friction effect on scratch-induced deformation in polymers. Polym Eng Sci 2013;53:1405–13. [12] Jardret V, Morel P. Viscoelastic effects on the scratch resistance of polymers: Relationship between mechanical properties and scratch properties at various temperatures. Prog Org Coat 2003;48:322–31. [13] Jardret V, Zahouania H, Loubet JL, Mathia TG. Understanding and quantification of elastic and plastic deformation during a scratch test. Wear 1998;218:8–14. [14] Jiang H, Browning RL, Hossain MM, Sue H-J, Fujiwara M. Quantitative evaluation of scratch visibility resistance of polymers. Appl Surf Sci 2010;256:6324–9. [15] Jiang H, Browning RL, Sue H-J. Understanding of scratch-induced damage mechanisms in polymers. Polymer 2009;50:4056–65. [16] Pelletier H, Durier A-L, Gauthier C, Schirrer R. Viscoelastic and elastic–plastic behaviors of amorphous polymeric surfaces during scratch. Tribol Int 2008;41:975–84. [17] Pelletier H, Gauthier C, Schirrer R. Strain and stress fields during scratch tests on amorphous polymers: Influence of the local friction. Tribol Lett 2008;32:109–16. [18] Pelletier H, Gauthier C, Schirrer R. Influence of the friction coefficient on the contact geometry during scratch onto amorphous polymers. Wear 2010;268:1157–69. [19] Taljat B, Pharr GM. Development of pile-up during spherical indentation of elastic-plastic solids. Int J Solids Struct 2004;41:3891–904. [20] Tang H, Martin DC. Near-surface deformation under scratches in polypropylene blends Part I Microscopic characterization of deformation. J Mater Sci 2003;38:803–15. [21] Xiang C, Sue H-J, Chu J, Coleman B. Scratch behavior and material property relationship in polymers. J Polym Sci Part B: Polym Phys 2001;39:47–59. [22] Al-Rub RKA, Faruk ANM. Dislocation-based model for predicting size-scale effects on the micro and nano indentation hardness of metallic materials. Int J Mater Struct Integr 2010;4:251–77. [23] Al-Rub RKA, Faruk ANM. Coupled interfacial energy and temperature effects on size-dependent yield strength and strain hardening of small metallic volumes. J Eng Mater Technol-T ASME 2011;133:011017. [24] Al-Rub RKA, Faruk ANM. Prediction of micro and nano indentation size effects from spherical indenters. Mech Adv Mater Struct 2012;19:119–28.
257
[25] ASTM D7027-05. Standard test method for evaluation of scratch resistance of polymeric coatings and plastics using an instrumented scratch machine. Pennsylvania: ASTM International; 2005. [26] Hossain MM. Quantitative modeling of polymer scratch behavior. College Station, Texas: Texas A&M University; 2013. [27] Hossain MM, Minkwitz R, Charoensirisomboon P, Sue H-J. Quantitative modeling of scratch-induced deformation in amorphous polymers. Polymer 2014;55:6152–66. [28] Bucknall CB, Ayre DS, Dijkstra DJ. Detection of rubber particle cavitation in toughened plastics using thermal contraction tests. Polymer 2000;41:5937–47. [29] Bucknall CB, Smith RR. Stress-whitening in high-impact polystyrenes. Polymer 1965;6:437–46. [30] Katime I, Quintana JR, Price C. Influence of the microstructural morphology on the mechanical properties of high-impact polystyrene. Mater Lett 1995;22:297–301. [31] Schmitt JA. Mechanism of reinforcement in rubber-modified polystyrene systems studied by use of a miniature dart drop test. J Appl Polym Sci 1968;12:533–46. [32] Browning RL, Jiang H, Moyse A, Sue H-J, Iseki Y, Ohtani K, Ijichi Y. Scratch behavior of soft thermoplastic olefins: effects of ethylene content and testing rate. J Mater Sci 2008;43:1357–65. [33] Moghbelli E, Browning RL, Boo W-J, Hahn SF, Feick LJE, Sue H-J. Effects of molecular weight and thermal history on scratch behavior of polypropylene thin sheets. Tribol Int 2008;41:425–33. [34] Moghbelli E, Sun L, Jiang H, Boo W-J, Sue H-J. Scratch behavior of epoxy nanocomposites containing α-zirconium phosphate and core-shell rubber particles. Polym Eng Sci 2009;49:483–90. [35] Liang Y-L, Sue H-J, Minkwitz R. Rubber content effect on scratch behavior in acrylonitrile-styrene-acrylate copolymers. J Appl Polym Sci 2012;126:1088–96. [36] Akay M, Ozden S. The influence of residual stresses on the mechanical and thermal properties of injection moulded ABS copolymer. J Mater Sci 1995;30:3358–68. [37] Liang Y-L, Moghbelli E, Sue H-J, Minkwitz R, Stark R. Effect of high temperature annealing on scratch behavior of acrylonitrile styrene acrylate copolymers. Polymer 2012;53:604–12. [38] Hamilton GM. Explicit equations for the stresses beneath a sliding spherical contact. Proc Inst Mech Eng Part C-J Mech Eng Sci 1983;197:53–9. [39] Hamilton GM, Goodman LE. The stress field created by a circular sliding contact. J Appl Mech 1966;33:371–6. [40] Liu J, Sue H-J, Thompson ZJ, Bates FS, Dettloff M, Jacob G, Verghese N, Pham H. Nanocavitation in self-assembled amphiphilic block copolymer-modified epoxy. Macromolecules 2008;41:7616–24. [41] Liu J, Sue H-J, Thompson ZJ, Bates FS, Dettloff M, Jacob G, Verghese N, Pham H. Strain rate effect on toughening of nano-sized PEP–PEO block copolymer modified epoxy. Acta Mater 2009;57:2691–701. [42] Liu J, Thompson ZJ, Sue H-J, Bates FS, Hillmyer MA, Dettloff M, Jacob G, Verghese N, Pham H. Toughening of epoxies with block copolymer micelles of wormlike morphology. Macromolecules 2010;43:7238–43. [43] Bhushan B. Introduction to tribology. 2nd ed.. New York: John Wiley & Sons Ltd; 2013. [44] Jiang H, Browning RL, Fincher J, Gasbarro A, Jones S, Sue H-J. Influence of surface roughness and contact load on friction coefficient and scratch behavior of thermoplastic olefins. Appl Surf Sci 2008;254:4494–9.
Contents lists available at ScienceDirect
Tribology International journal homepage: www.elsevier.com/locate/triboint
Effect of thermal history on scratch behavior of multi-phase styrenic-based copolymers Mohammad Motaher Hossain a, Eike Jahnke b, Philipp Boeckmann c, Svetlana Guriyanova c, Rolf Minkwitz c, Hung-Jue Sue a,n a
Polymer Technology Center Department of Materials Science and Engineering, Texas A&M University, College Station, TX 77843, USA Styrolution Group GmbH, 60325 Frankfurt, Germany c BASF SE, Advanced Materials & Systems Research, 67056 Ludwigshafen, Germany b
art ic l e i nf o
a b s t r a c t
Article history: Received 13 November 2015 Received in revised form 23 March 2016 Accepted 24 March 2016 Available online 31 March 2016
Effect of thermal history on scratch performance of acrylonitrile styrene acrylate (ASA) and acrylonitrile butadiene styrene (ABS) model systems was studied. ASAs with rubber particle sizes of 100 nm and 1 mm, respectively, and ABS with rubber particle size of 100 nm were investigated. Linearly increasing normal load scratch tests were performed according to the ASTM-D7027/ISO-19252 standard. An improvement in scratch resistance is observed in ASAs when annealed at 140 °C whereas a noticeable drop in scratch resistance is found in ABS under the same condition. Various microscopic analyses were conducted to determine the corresponding deformation mechanisms for the observed dependence of scratch performance on annealing. Implication of the present findings for design of scratch resistant rubber toughened polymers is discussed. & 2016 Elsevier Ltd. All rights reserved.
Keywords: Rubber modification Heat treatment Scratch resistance Damage mechanisms
1. Introduction Scratch behavior of multi-phase polymeric systems is gaining significant interests due to its increasingly wide usage in engineering, and structural and durable goods applications. However, in-depth understanding of scratch-induced surface deformation in multi-phase polymeric systems is quite challenging since the presence of micro- and nano-phase domains near the surface can greatly influence the scratch deformation and damage [1]. It is not straightforward to relate the surface deformation and damage phenomena to bulk mechanical properties in multi-phase polymeric systems. For instance, rubber addition to the matrix improves toughness and ductility but simultaneously reduces modulus and strength [2,3], which complicates the fundamental understanding of their influences on scratch behavior. For homogeneous single-phase polymers, researchers have shown some success in correlating material and surface properties to the evolution of different scratch-induced surface damage features [4–24]. Using the ASTM D7027/ISO 19252 scratch test standard [25], it has been shown that an increase in tensile strength delays the onset of microcrack formation during scratching in styrene-acrylonitrile (SAN) random copolymers [5]. Bucaille et al. n
Corresponding author. Tel.: þ 1 979 845 5024; fax: þ 1 979 845 3081. E-mail address: [email protected] (H.-J. Sue).
http://dx.doi.org/10.1016/j.triboint.2016.03.026 0301-679X/& 2016 Elsevier Ltd. All rights reserved.
[7] showed that a more pronounced strain hardening can improve scratch resistance, defined as the resistance to surface deformation and damage due to sliding indentation of a rigid asperity, since it restricts large-scale plastic deformation. Using the finite element method (FEM) parametric study and experimental observation, it has been shown that, in addition to the coefficient of friction, yield stress, strain at stress recovery and strain hardening slope in compression are the most important material parameters that determine the development of scratch groove during scratching [9–11]. Tensile behavior has little influence on scratch groove formation but correlates well with microcrack formation during scratching [9–11]. In a recent study [26,27], it has been shown that by knowing the material behavior and surface characteristics, the scratch behavior of polymers can be quantitatively predicted with reasonable accuracy. For multi-phase polymeric systems, understanding and analysis of the scratch behavior is rather complex. In-depth understanding of the localized deformation and damage due to interaction between the micro-/nano-sized particles and matrix is needed. Introduction of a rubber phase in a polymer matrix has shown to significantly alter the stress state near the rubber particles [28–31], which could ultimately change the deformation mechanisms. Browning et al. [32] showed that the crystallinity of the ethylene segment and the internal morphology of the ethylene-propylene rubber (EPR) phase significantly affect the
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
scratch behavior of soft thermoplastic olefins (TPOs). In another study [33], it has been shown that the degree of surface crystallinity, which can be modified by varying the processing conditions, can significantly influence the scratch behavior of polypropylene (PP). It was further shown that [34] improvements in different bulk mechanical properties by adding synthetic clay nanoplatelets or core-shell rubber (CSR) nanoparticles in epoxy do not necessarily enhance the scratch resistance. Liang et al. [35] showed that the scratch resistance of rubber-modified acrylonitrile styrene acrylate (ASA) system deteriorates with increasing rubber content due to reduction in tensile and compressive properties. In a more recent study [1], Hossain et al. reported a reduction in scratch resistance in ASA containing 100 nm rubber particles when compared to ASA containing 1 mm rubber particles. It was further shown that rubber type has significant influence on scratch behavior as acrylonitrile butadiene styrene (ABS) exhibits a higher scratch resistance when compared to ASA with comparable rubber particle size. Since the bulk mechanical behaviors in tension and compression of the model systems are similar, they were not so useful in explaining the observed difference in scratch resistance in the rubber-modified systems. It was argued [1] that the rubber particle type and size can alter the damping characteristics of the SAN matrix, hence the frictional behavior to cause variation in their scratch behavior. Thermal history has long been shown to significantly affect the physical and mechanical properties of different polymers. Variation in processing condition can induce and alter skin-core characteristics and significantly affect the final morphology, especially on sample surfaces. However, for styrenic-based copolymers, it has been shown that the processing conditions have insignificant effect on mechanical properties [36,37]. In particular, tensile yield stress and Charpy impact strength have been shown to remain the same upon variation in heat treatment procedure. It was also shown that [37] high temperature annealing can improve the scratch resistance without influencing the bulk properties in rubber-modified styrene-acrylonitrile (SAN) systems. The present study focuses on understanding how the thermal history influences the scratch behavior of styrenic-based Table 1 Material information of model systems.
ASA100 ABS100 ASA1000
Rubber type
Structure type
Rubber particle size
ASA ABS ASA
SAN grafted PBA rubber SAN grafted PBD rubber SAN grafted PBA rubber
E100 nm E100 nm E1 mm
249
copolymers. The model systems, acrylonitrile styrene acrylate (ASA) and acrylonitrile butadiene styrene (ABS) are composed of SAN matrix, and acrylic and butadiene rubbers, respectively. Extensive analysis has been performed to understand the observed differences in scratch resistance and evolution of different scratchinduced surface deformation mechanisms. Implications of the present study for designing scratch resistant rubber-modified polymers are discussed.
2. Experimental 2.1. Materials The ASA and ABS copolymers utilized in this study were provided by Styrolution Group GmbH (Frankfurt, Germany). The ASA systems consist of a random copolymer SAN matrix and grafted polybutyl-acrylate (PBA) rubber particles with an average nominal diameter of 100 nm and 1 mm, respectively. Size distribution of the rubber particles in the model systems can be found elsewhere [1]. In the SAN phase, the acrylonitrile content was controlled to be at 35 wt%, and the weight-average molecular weight (Mw) of SAN was chosen to be 104 kg/mol [1] with a polydispersity (PDI) of 3.7. The ABS system consisted of a random copolymer SAN matrix and grafted polybutadiene (PBD) rubber particles with an average nominal diameter of approximately 100 nm. Within the SAN phase the same material as described for ASA was used. The rubber concentration is 30 wt% for all the systems investigated. Test specimens were fabricated by injection molding with a fan gate design to spread and slow the melt as it enters the mold cavity to ensure near-uniform molecular orientation across the width of the plaques. The plaques were 150 mm 150 mm in rectangular shape and 6 mm in thickness. Nomenclature and physical characteristics of each system are listed in Table 1. 2.2. Heat treatment Upon receipt, all the plaques were dried in a vacuum oven for 12 h at 80 °C with a vacuum pressure of 30 mm Hg. The drying process was conducted by sandwiching the plaques in between two smooth glass plates (Surface roughness, Ra ¼17.0 nm, Rq ¼22.3 nm). An external weight leading to stress of about 10 kPa was applied to the sandwiched samples during heating and cooling to prevent the potential for uneven residual stress relief or warping of the plaques. After drying, the high temperature annealing (HTA) process was carried out on one of the two dry reference plaques for each system studied. The HTA process was conducted at 140 °C, which is about
Fig. 1. AFM micrographs of the ASA100 cross-section in the sub-surface, (a) dry reference sample and (b) HTA sample.
250
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
Fig. 2. AFM micrographs of the ABS100 cross-section in the sub-surface, (a) dry reference sample and (b) HTA sample.
Fig. 3. AFM micrographs of the ASA1000 cross-section in the sub-surface, (a) dry reference sample and (b) HTA sample.
30 °C higher than the measured Tg [1], for 2 h under vacuum pressure of 30 mm Hg. Afterwards, the samples were slowly cooled to ambient temperature at 1.8 °C/min. The heat treatment procedure is free from any polymer degradation due to oxidation. All the samples (both Dry reference and HTA) were kept in a desiccator before the scratch tests to minimize moisture adsorption from the atmosphere. 2.3. Scratch test Scratch tests were carried out according to the ASTM D7027 standard [25] by using a progressive normal load range of 1–70 N at a constant scratch speed of 1 mm/s for a length of 100 mm. A stainless steel spherical scratch tip with diameter of 1 mm was used. A minimum of five scratch tests were performed at room temperature on each plaque in the melt flow direction of the sample, in accordance with the ASTM standard to allow for controlled testing and analysis. 2.4. Analysis of scratch-induced damage Analysis of scratch-induced surface damage was carried out 72 h after the completion of scratch tests to allow for sufficient viscoelastic recovery. A Keyences VK9700 violet laser scanning confocal microscope (VLSCM) was used for high-resolution surface topographical analysis. The scratch depth and shoulder height at different locations on the scratch path were measured using the associated VK Analyzer software. The critical normal loads for the onsets of cracking and
plowing were measured, as well. The onset of cracking was determined at the point where periodic repetition of the microcracks in or around the center of the scratch groove started by carefully observing the whole scratch under the microscope. The height profile along the scratch path, generated by the VK Analyzer software, was also analyzed along with the optical images to precisely locate the onset of microcrack formation. Surface roughness of the unscratched sample was also measured to study the effect of thermal history on surface morphology. When reporting the data, error bars in the plots indicate one standard deviation (or 1σ) of the mean value.
2.5. Dynamic mechanical analysis (DMA) Dynamic mechanical analysis in torsional mode was performed on a TA Instruments ARES G2 Rheometer. Dry samples with nominal dimensions of 30 mm 10 mm 3 mm were prepared for the DMA. The samples were cut by an Isomet 1000 Precision Saw and the surfaces of the specimens were polished using 4000 grit silicon carbide polishing paper until a surface roughness of 1 mm or less has been achieved. Strain amplitude of 0.05% was chosen for the analysis. The samples were tested from 135 °C to 180 °C with a temperature increment of 3 °C/min at a fixed frequency of 1 Hz. Storage modulus (G0 ) and loss modulus (G00 ) information was recorded during thetest and corresponding loss 00
tangent value ( tan δ) was calculated tan δ ¼ GG0 .
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
Fig. 4. Plot of a) scratch depth and b) shoulder height, as a function of scratch normal load in ASA100.
251
Fig. 5. Plot of a) scratch depth and b) shoulder height, as a function of scratch normal load in ASA1000.
2.6. Atomic force microscopy (AFM) The characterization of sample morphology on the crosssection in the sub-surface using AFM (Dimension, Veeco) was carried out by BASF SE (Ludwigshafen, Germany). The AFM imaging and analysis were performed in Tapping Mode. The cantilever was excited at its resonance frequency and moved at a defined height over the sample surface. Thus, interaction forces cause a variation in the oscillation amplitude of the lever. Furthermore, a phase shift between the excitation and response oscillation can be induced by different material properties such as stiffness. In addition to the height, this effect provides material contrast information of the sample (Phase images). An Olympus silicon tip with cantilever spring constant of 40 N/m was used. The smooth cross-section surfaces of the samples were prepared using microtome technique at 80 °C. The AFM measurements were carried out at room temperature. 2.7. AFM peak force quantitative nanomechanical property mapping (AFM PF-QNM) Peak force quantitative nanomechanical property mapping (PFQNMs) allows quantitative nanomechanical mapping of material properties, including modulus, adhesion, deformation and dissipation, while simultaneously imaging sample topography at high resolution. The surface morphology of the model systems was characterized in-house using AFM (Bruker Dimension Icon AFM) in peak force Tapping Mode. ScanAsysts, which uses algorithms to automatically and continuously monitor image quality and make
Fig. 6. Critical normal loads for onsets of microcracking in the model systems.
the appropriate parameter adjustments, was used for the imaging purpose. Material properties, such as, modulus and deformation, was mapped on the surface of the model systems. The cantilever spring constant of 40 N/m was used for the AFM measurements, carried out at room temperature.
3. Results and discussion AFM micrographs of the dry reference and annealed samples under cross-sectional view in the sub-surface (ca. 50 mm below the
252
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
Fig. 7. Critical normal loads for onsets of plowing in the model systems.
surface) are shown in Figs. 1–3. In both ASA100 and ABS100, the rubber particle size is approximately 100 nm; whereas in ASA1000, the rubber is 1 mm in size. It can be observed that the thermal history has negligible effect on the shape, size and orientation of the rubber particles in the sub-surface. Figs. 4 and 5 show the plots of evolution of residual scratch depth and shoulder height of the scratch groove as a function of scratch normal load obtained via VLSCM for ASA100 and ASA1000, respectively. The numbers 1 to 5 in the plots denote five individual scratch measurements on the same sample. As discussed, the scratch damages were scanned 72 h after the scratch tests to allow for viscoelastic recovery. All measurements were conducted from low load towards high load up to the occurrence of well-developed cracks as they can obscure the scratch depth measurement. Since ABS100 shows onset of crack formation at very low load, to be discussed later, the scratch depth and shoulder height of the scratch groove was not measured for ABS100. As can be seen in Figs. 4 and 5, the scratch depth and shoulder height are similar despite the difference in thermal history for both ASAs. Thus, the difference in heat treatment has negligible influence on the development of scratch groove along the scratch path. Critical normal loads for onsets of microcracking and plowing along the scratch path for the model systems are shown in Figs. 6 and 7, respectively. All the rubber-modified systems show deterioration of scratch resistance with earlier onsets of damage transitions when compared to the neat system we have studied earlier (For neat system, onset of microcracking: 45 N, onset of plowing: 68 N) [5]. For ASAs, high temperature annealing is observed to improve the scratch resistance by delaying the onset of microcracking. The onset of plowing for ASA1000 is also delayed due to annealing. On the contrary, high temperature annealing has induced an earlier onset of microcrack formation in ABS100 although it slightly delayed the onset of plowing. Thus, the effect of thermal history on scratch resistance is greatly influenced by these two rubber types in the rubber-modified systems. It should be noted that the difference in onset of damage transitions between dry ASA100 and ABS100 model systems is insignificant compared to our previous study [1]. However, the drying procedures in both cases are different. Fig. 8 shows the comparison of representative scratch coefficient of friction (SCOF ¼ FFnt ; where Ft is the tangential force and Fn is the normal load) curves obtained during experiment for the model systems. The SCOF consists of two terms: a conventional frictional term, i.e., coefficient of friction (COF) due to adhesion and an additional term due to material deformation. As can be seen in the figure, dry reference systems show higher SCOF compared to their HTA counterparts for ASAs at low
Fig. 8. Representative scratch coefficient of friction (SCOF) curves for a) ASA100, b) ABS100, and c) ASA1000.
normal load regime. Since the residual scratch depth remains essentially the same for both dry reference and HTA samples in ASAs (Figs. 4 and 5) and it has been shown that the heat treatment process has insignificant influence on bulk mechanical properties [37], the discrepancy in the SCOF can possibly be due to the difference in coefficient of friction due to adhesion (i.e., COF). It has been shown that the onset of microcrack formation depends on COF as it intensifies the tensile stress state on the surface behind the scratch tip [38,39]. Thus, the difference in COF can be considered responsible for earlier onset of microcrack formation in the dry reference samples in ASAs. For
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
253
Fig. 9. Optical image at onset of microcrack formation for a) ASA100, b) ABS100, and c) ASA1000. (Scratch direction: from left to right; arrow indicates the onset point on the track).
Fig. 10. Optical image of crack patterns for a) ASA100, and b) ASA1000. (Scratch direction: from left to right).
254
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
Fig. 11. Transition of crack patterns in ABS100 a) dry reference, and b) HTA. (Scratch direction: from left to right).
Fig. 12. Loss tangent (tan δ) of the model systems as a function of temperature below T g .
ABS100, the SCOF remains essentially the same before the onset load for microcracking in dry reference and HTA samples, unable to explain the observed difference in onset of microcracking. Optical micrographs of the onset of scratch-induced microcrack formation in each model system are shown in Fig. 9. It can be seen that the microcrack patterns differ among the rubber-modified systems. Thermal history has no influence on the crack pattern. For ASAs, the microcrack formation around the onset appears to be caused by tensile fracture as denoted by the opening type of cracks [1,15]. Whereas, for ABS100, the damage pattern in the initial stage can be considered shear dominated mixed-mode in nature as shown by the peeling kind of damage [1]. As scratch normal
Fig. 13. Surface roughness of the model systems.
load increases, the microcracks develop into larger cracks for both ASAs (Fig. 10). However, for ABS100, the damage pattern eventually turns from mixed-mode to tensile dominated fracture (Fig. 11). The damage patterns observed in the dry reference system in this study are consistent with the literature [1]. The reason for the difference in the damage pattern between ASAs and ABS can be due to the difference in miscibility between the rubber particles and the matrix. DMA study on the model systems was conducted to support the conjecture. DMA study of the model systems was carried out to probe their viscoelastic behavior. The tan δ plots in semi-logarithmic scale before the glass transition temperature (Tg) are shown in Fig. 12. Tg, which is defined as the temperature at the maximum peak in tan δ curve, does not change with the presence of rubber particles
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
255
Fig. 14. Modulus mapping on the surface using peak force quantitative nanomechanical property mapping (PF-QNM) for ASA100 a) dry reference, b) HTA, and c) modulus distribution. (Darker regions indicate softer material, brighter regions indicate harder material).
Fig. 15. Modulus mapping on the surface using peak force quantitative nanomechanical property mapping (PF-QNM) for ABS100 a) dry reference, b) HTA, and c) modulus distribution. (Darker regions indicate softer material, brighter regions indicate harder material).
Fig. 16. Modulus mapping on the surface using peak force quantitative nanomechanical property mapping (PF-QNM) for ASA1000 a) dry reference, b) HTA, and c) modulus distribution. (Darker regions indicate softer material, brighter regions indicate harder material).
[1]. The secondary transition denoting the Tg of the rubber particles can be identified in the plots ( 90 °C for ABS and 45 °C for ASAs). Introduction of rubber particles in SAN matrix has altered the damping characteristics of the matrix. ASA100 shows the most pronounced damping behavior change at temperatures above the T g of the butyl-acrylate rubber, i.e., 45 °C. This suggests that the 100 nm size butyl-acrylate rubber influences the damping behavior of the SAN matrix most significantly in the model systems investigated. The T g of the butyl-acrylate rubber also appears to have shifted to a higher temperature in ASA100 compared to
ASA1000, suggesting good interaction between the 100 nm butylacrylate rubber and SAN matrix. On the other hand, for ASA1000, the butyl-acrylate rubber does not seem to exert a similar degree of influence in damping behavior of SAN matrix. Interestingly, the 100 nm size butadiene rubber in ABS100 does not show the same influence on SAN damping behavior as ASA100. This suggests that butadiene rubber is less miscible with SAN [1,40–42]. As a result, the interface between the rubber particles and matrix can be considered sharp and the adhesion between the rubber particles and the matrix is likely to be weak in ABS100 compared to those of the ASAs. These differences can be responsible for the observed
256
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
variation in scratch damage patterns. It should be noted that, no observable degradation is detected due to high temperature annealing process shown in one of the previous studies [37]. Surface roughness may influence the effective contact area during sliding contact, which alters the COF [43,44], and, consequently the scratch behavior [11]. However, if the surface roughness is sufficiently low and the applied stress is high enough to induce large scale plastic deformation through the roughness layer, the roughness effect on COF becomes negligible [43,44]. The similarities in residual scratch depth and shoulder height between dry reference and HTA systems in ASAs (Figs. 4 and 5) show that indeed the change in surface roughness due to annealing process has insignificant influence on scratch-induced deformation and damage. Although the amount of surface roughness change due to annealing has negligible influence on scratch behavior, the topographical change can provide useful insights on molecular relaxation and strain recovery due to annealing. Fig. 13 shows the RMS roughness values of the dry reference and HTA unscratched samples, measured on an area of 525 mm 700 mm (at 20 magnification in VLSCM). The area chosen for surface roughness measurement is comparable with the scratch tip geometry and deformation of interest. As can be seen in the figure, the surface roughness increases considerably due to the annealing process for the ASAs. The increase in surface roughness is more pronounced in ASA100 than in ASA1000. Interestingly, for ABS100, the surface roughness change is minimal. The change in surface roughness can be due to relaxation of the highly stretched molecules near the surface which causes rubber particles to be more exposed to the surface during the annealing process [37], leading to a significant increase in surface roughness. The validity of this conjecture will be explored further in the peak force quantitative nanomechanical modulus mapping on the surface. The difference in the degrees of surface roughness change between the ASAs and ABS is probably due to a more coordinated stretching between SAN molecules and butyl-acrylate rubber particles than between SAN and butadiene rubber particles upon injection molding. Figs. 14–16 show the modulus mapping on the surface obtained from peak force quantitative nanomechanical mapping (PF-QNM) for ASA100, ABS100 and ASA1000, respectively. The dark regions on the images indicate softer or lower modulus material whereas the brighter regions indicate the opposite. The rubber particle size and shape on the surface can be observed from the modulus mapping. The modulus distribution on the surface is calculated by normalizing the matrix modulus obtained from PF-QNM against the neat matrix modulus provided by BASF SE. For ASAs, the modulus distribution is shifted towards lower modulus for rubber particles in HTA sample compared to the dry reference counterpart. This shift is most prominent in ASA1000 system indicating that there might be a coating of matrix material on top of the rubber particles in the dry reference system, which was then partially/fully removed upon high temperature annealing. This phenomenon can be due to the retraction of SAN matrix surrounding the rubber particles during the annealing process. The rubber particles have also recovered some of its injection molding induced orientation. It is noted that the matrix modulus on the surface becomes more distributed in the HTA system compared to that of dry reference sample, suggesting that the annealing process has helped reduce stress concentration at the rubber and matrix interface, which might contribute to a delay in onset of microcracking in the HTA systems in ASAs. For ABS100, however, the mechanism of microcrack formation is completely different. The damage pattern is observed to be mixed-mode fracture which eventually turns into tensile dominated fracture at higher normal loads (Figs. 9–11). As can be seen in Fig. 15, no significant difference can be found between the HTA sample and dry reference system in terms of modulus distribution
on the surface. Since the onset of microcrack formation in ABS100 is due to mixed-mode fracture, the interfacial adhesion between rubber particles and matrix are expected to play a major role. The annealing process can induce a greater relaxation of rubber particles and matrix molecules in ABS100 leading to sharper interface and weaker adhesion compared to the dry reference counterpart. As a result, HTA sample shows an earlier onset of microcracking compared to dry reference system in ABS100 as shown in Fig. 6. The present study suggests that the scratch behavior of rubbermodified systems depends on thermal history. The influence of thermal history on scratch behavior depends on the rubber type. The miscibility of the rubber particles with the matrix and corresponding viscoelastic response can be considered responsible for the observed difference in damage pattern formed. Change in surface morphology due to high temperature annealing can be the reason for observed difference in scratch resistance of the model systems investigated. More in-depth study is required to further investigate the complex stress and strain fields developed around the rubber particles in rubber-modified systems.
4. Conclusions The effect of thermal history was studied on scratch performance of two types of styrenic-based copolymers with variation in rubber type and size. Linearly increasing normal load scratch tests were performed according to the ASTM standard for scratch testing to characterize their scratch performance. Significant change in surface morphology has been observed in the ASAs due to possible retraction of SAN matrix surrounding the rubber particles. On the contrary, ABS shows less retraction around the rubber particles. An improvement in scratch resistance is observed in the ASAs when annealed at high temperature whereas a noticeable drop in scratch resistance is found in ABS upon annealing. Thus, the influence of thermal history on scratch performance of rubber-modified systems has shown to be dependent on the rubber type. Due to multiphase nature of the systems studied, scratch-induced deformation and its relation to the thermal history is rather complicated in the rubber-modified systems, and needs further comprehensive understanding. Additional studies on the systems containing rubber particle sizes between 100–1000 nm and different rubber types could further elucidate the effect of thermal history on scratch behavior of rubber modified systems.
Acknowledgments The authors would like to acknowledge the financial support and valuable insights provided by Styrolution Group GmbH in this research endeavor. Special thanks are also given for providing the model polymer systems.
References [1] Hossain MM, Moghbelli E, Jahnke E, Boeckmann P, Guriyanova S, Sander R, Minkwitz R, Sue H-J. Rubber particle size and type effects on scratch behavior of styrenic-based copolymers. Polymer 2015;63:71–81. [2] Kim H, Keskkula H, Paul DR. Toughening of SAN copolymers by an SAN emulsion grafted rubber. Polymer 1990;31:869–76. [3] Steenbrink AC, Litvinov VM, Gaymans RJ. Toughening of SAN with acrylic coreshell rubber particles: particle size effect or cross-link density? Polymer 1998;39:4817–25. [4] Browning RL, Lim GT, Moyse A, Sun L, Sue H-J. Effects of slip agent and talc surface-treatment on the scratch behavior of thermoplastic olefins. Polym Eng Sci 2006;46:601–8.
M.M. Hossain et al. / Tribology International 99 (2016) 248–257
[5] Browning RL, Sue H-J, Minkwitz R, Charoensirisomboon P. Effects of acrylonitrile content and molecular weight on the scratch behavior of styreneacrylonitrile random copolymers. Polym Eng Sci 2011;51:2282–94. [6] Bucaille JL, Felder E, Hochstetter G. Mechanical analysis of the scratch test on elastic and perfectly plastic materials with the three-dimensional finite element modeling. Wear 2001;249:422–32. [7] Bucaille JL, Gauthier C, Felder E, Schirrer R. The influence of strain hardening of polymers on the piling-up phenomenon in scratch tests: Experiments and numerical modelling. Wear 2006;260:803–14. [8] Chu J, Rumao L, Coleman B. Scratch and mar resistance of filled polypropylene materials. Polym Eng Sci 1998;38:1906–14. [9] Hossain MM, Browning RL, Minkwitz R, Sue H-J. Effect of asymmetric constitutive behavior on scratch-induced deformation of polymers. Tribol Lett 2012;47:113–22. [10] Hossain MM, Jiang H, Sue H-J. Effect of constitutive behavior on scratch visibility resistance of polymers-A finite element method parametric study. Wear 2011;270:751–9. [11] Hossain MM, Minkwitz R, Sue H-J. Minimization of surface friction effect on scratch-induced deformation in polymers. Polym Eng Sci 2013;53:1405–13. [12] Jardret V, Morel P. Viscoelastic effects on the scratch resistance of polymers: Relationship between mechanical properties and scratch properties at various temperatures. Prog Org Coat 2003;48:322–31. [13] Jardret V, Zahouania H, Loubet JL, Mathia TG. Understanding and quantification of elastic and plastic deformation during a scratch test. Wear 1998;218:8–14. [14] Jiang H, Browning RL, Hossain MM, Sue H-J, Fujiwara M. Quantitative evaluation of scratch visibility resistance of polymers. Appl Surf Sci 2010;256:6324–9. [15] Jiang H, Browning RL, Sue H-J. Understanding of scratch-induced damage mechanisms in polymers. Polymer 2009;50:4056–65. [16] Pelletier H, Durier A-L, Gauthier C, Schirrer R. Viscoelastic and elastic–plastic behaviors of amorphous polymeric surfaces during scratch. Tribol Int 2008;41:975–84. [17] Pelletier H, Gauthier C, Schirrer R. Strain and stress fields during scratch tests on amorphous polymers: Influence of the local friction. Tribol Lett 2008;32:109–16. [18] Pelletier H, Gauthier C, Schirrer R. Influence of the friction coefficient on the contact geometry during scratch onto amorphous polymers. Wear 2010;268:1157–69. [19] Taljat B, Pharr GM. Development of pile-up during spherical indentation of elastic-plastic solids. Int J Solids Struct 2004;41:3891–904. [20] Tang H, Martin DC. Near-surface deformation under scratches in polypropylene blends Part I Microscopic characterization of deformation. J Mater Sci 2003;38:803–15. [21] Xiang C, Sue H-J, Chu J, Coleman B. Scratch behavior and material property relationship in polymers. J Polym Sci Part B: Polym Phys 2001;39:47–59. [22] Al-Rub RKA, Faruk ANM. Dislocation-based model for predicting size-scale effects on the micro and nano indentation hardness of metallic materials. Int J Mater Struct Integr 2010;4:251–77. [23] Al-Rub RKA, Faruk ANM. Coupled interfacial energy and temperature effects on size-dependent yield strength and strain hardening of small metallic volumes. J Eng Mater Technol-T ASME 2011;133:011017. [24] Al-Rub RKA, Faruk ANM. Prediction of micro and nano indentation size effects from spherical indenters. Mech Adv Mater Struct 2012;19:119–28.
257
[25] ASTM D7027-05. Standard test method for evaluation of scratch resistance of polymeric coatings and plastics using an instrumented scratch machine. Pennsylvania: ASTM International; 2005. [26] Hossain MM. Quantitative modeling of polymer scratch behavior. College Station, Texas: Texas A&M University; 2013. [27] Hossain MM, Minkwitz R, Charoensirisomboon P, Sue H-J. Quantitative modeling of scratch-induced deformation in amorphous polymers. Polymer 2014;55:6152–66. [28] Bucknall CB, Ayre DS, Dijkstra DJ. Detection of rubber particle cavitation in toughened plastics using thermal contraction tests. Polymer 2000;41:5937–47. [29] Bucknall CB, Smith RR. Stress-whitening in high-impact polystyrenes. Polymer 1965;6:437–46. [30] Katime I, Quintana JR, Price C. Influence of the microstructural morphology on the mechanical properties of high-impact polystyrene. Mater Lett 1995;22:297–301. [31] Schmitt JA. Mechanism of reinforcement in rubber-modified polystyrene systems studied by use of a miniature dart drop test. J Appl Polym Sci 1968;12:533–46. [32] Browning RL, Jiang H, Moyse A, Sue H-J, Iseki Y, Ohtani K, Ijichi Y. Scratch behavior of soft thermoplastic olefins: effects of ethylene content and testing rate. J Mater Sci 2008;43:1357–65. [33] Moghbelli E, Browning RL, Boo W-J, Hahn SF, Feick LJE, Sue H-J. Effects of molecular weight and thermal history on scratch behavior of polypropylene thin sheets. Tribol Int 2008;41:425–33. [34] Moghbelli E, Sun L, Jiang H, Boo W-J, Sue H-J. Scratch behavior of epoxy nanocomposites containing α-zirconium phosphate and core-shell rubber particles. Polym Eng Sci 2009;49:483–90. [35] Liang Y-L, Sue H-J, Minkwitz R. Rubber content effect on scratch behavior in acrylonitrile-styrene-acrylate copolymers. J Appl Polym Sci 2012;126:1088–96. [36] Akay M, Ozden S. The influence of residual stresses on the mechanical and thermal properties of injection moulded ABS copolymer. J Mater Sci 1995;30:3358–68. [37] Liang Y-L, Moghbelli E, Sue H-J, Minkwitz R, Stark R. Effect of high temperature annealing on scratch behavior of acrylonitrile styrene acrylate copolymers. Polymer 2012;53:604–12. [38] Hamilton GM. Explicit equations for the stresses beneath a sliding spherical contact. Proc Inst Mech Eng Part C-J Mech Eng Sci 1983;197:53–9. [39] Hamilton GM, Goodman LE. The stress field created by a circular sliding contact. J Appl Mech 1966;33:371–6. [40] Liu J, Sue H-J, Thompson ZJ, Bates FS, Dettloff M, Jacob G, Verghese N, Pham H. Nanocavitation in self-assembled amphiphilic block copolymer-modified epoxy. Macromolecules 2008;41:7616–24. [41] Liu J, Sue H-J, Thompson ZJ, Bates FS, Dettloff M, Jacob G, Verghese N, Pham H. Strain rate effect on toughening of nano-sized PEP–PEO block copolymer modified epoxy. Acta Mater 2009;57:2691–701. [42] Liu J, Thompson ZJ, Sue H-J, Bates FS, Hillmyer MA, Dettloff M, Jacob G, Verghese N, Pham H. Toughening of epoxies with block copolymer micelles of wormlike morphology. Macromolecules 2010;43:7238–43. [43] Bhushan B. Introduction to tribology. 2nd ed.. New York: John Wiley & Sons Ltd; 2013. [44] Jiang H, Browning RL, Fincher J, Gasbarro A, Jones S, Sue H-J. Influence of surface roughness and contact load on friction coefficient and scratch behavior of thermoplastic olefins. Appl Surf Sci 2008;254:4494–9.