Available online at www.sciencedirect.

com

Journal of the European Ceramic Society 33 (2013) 23592364

Tribological properties of Si3N4graphene nanocomposites

Pavol Hvizdo s a, , Jn Dusza a , Csaba Balzsi b
b

Institute of Materials Research, Slovak Academy of Sciences, Watsonova 47, 04353 Ko sice, Slovakia Institute for Technical Physics and Materials Science, Research Centre for Natural Sciences, Hungarian Academy of Sciences, P.O. Box 49, Budapest H-1525, Hungary Available online 1 May 2013

Abstract Mechanical and tribological properties of nanocomposites with silicon nitride matrix with addition of 1 and 3 wt% of various types of graphene platelets were studied. The wear behavior was observed by means of the ball-on-disk technique with a silicon nitride ball used as the tribological counterpart at room temperature in dry sliding. Coefcient of friction and specic wear rates were calculated and related to the damage mechanisms observed in the wear tracks. The measured properties were then assessed with respect to the type and volume fraction of the graphene additives. It is shown that addition of such amounts of carbon phases does not lower the coefcient of friction. Graphene platelets seem to be integrated into the matrix very strongly and they do not participate in lubricating processes. The best performance offers materials with 3 wt% of larger sized graphene, which have the highest wear resistance. 2013 Elsevier Ltd. All rights reserved.
Keywords: Si3 N4 ; Graphene; Carbon nanophases; Friction; Wear

1. Introduction Thanks to their exceptional mechanical and physical properties carbon nanomaterials have been proposed as strengthening ller for composites, or as specic functional element, for example for nanoelectronic devices. In the last few years new ceramic based nanocomposites have been developed with two aims: to improve mechanical properties of ceramic materials by the carbon nanotubes and carbon nanobres, and to develop ceramics with better functional properties using ceramic composites with carbon nanotubes/nanobres.17 Silicon nitride is a promising structural ceramic material that was developed in a search for high strength and high toughness ceramics that could replace metals in advanced turbine and reciprocating engines to give higher operating temperatures and efciencies. There have been explored many ways how to improve the fracture toughness and lower the brittleness of these materials, as well as how to reduce friction and enhance the wear resistance.8,9 Recently, in other types of ceramic matrices some success was achieved by adding carbon nanotubes (CNT) and graphene. In zirconia based systems a number of studies reported

Corresponding author. Tel.: +421 55 792 2464; fax: +421 55 792 2408. E-mail addresses: phvizdos@imr.saske.sk, hvizdosp@yahoo.com (P. Hvizdo s). 0955-2219/$ see front matter 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jeurceramsoc.2013.03.035

that very little addition of CNT or carbon nanobers (12 wt%) is sufcient to lower friction dramatically.10,11 Moderate success was achieved in aluminaCNT systems.12,13 However, to date not many studies of silicon nitride based materials exist, particularly about their friction and wear properties. Jones et al.14 tested silicon nitride with carbon additives in several modes, lubricated and in dry sliding, also under water conditions and under severe loading. They found that particularly in non-self-mated wear (against steel) small amount of carbon bers reduced the coefcient of friction (COF) by over 75%. Gonzalez-Julian et al.15 reported for silicon nitride with multi-walled CNTs (amounts up to 8.6 vol%) against Si3 N4 balls under lubrication in isooctane reduction of COF and wear by 40% and 80%, respectively. Hvizdo s et al.16 compared several ceramic matrix composites reinforced with CNTs and found that in silicon nitride higher amounts of CNTs are necessary to reduce COF. It had been recognized in these studies that good dispersion of CNTs which have a strong tendency to clustering remains a difcult challenge. The attention then turned also to graphene. To date, work on silicon nitride with graphene llers has been intensifying. Pfeifer et al.17 performed some introductory experiments with an experimental multilayer graphene. Kvetkov et al.18 showed toughening effect of various types of graphene materials introduced into silicon nitride matrix. Balzsi19,20 studied various CNT and multilayer graphene containing silicon nitride composites. Walker et al.21 reported signicant

2360

P. Hvizdo s et al. / Journal of the European Ceramic Society 33 (2013) 23592364 Table 1 Composition of experimental materials and their designation. Material Additives, wt% C SN SN-15 SN-125 SN-1A SN-1M SN-35 SN-325 SN-3A SN-3M
a

improvement of fracture toughness in graphene containing ceramic nanocomposites where they identied novel toughening micromechanisms. Ramirez et al.22,23 explored enhancement of electrical properties of graphene/silicon nitride composites. Those works report improvement of some electrical and mechanical properties, comment on some inconclusive results for wear of Si3 N4 CNT, but have not studied wear of Si3 N4 graphene. The aim of the present contribution is to investigate the inuence of the addition of various kinds of graphene nanoplatelets on the tribological properties of Si3 N4 based composites (COF and wear), observe and identify wear damage micromechanisms. 2. Materials and methods Several different types of materials were prepared. The basis was the reference material made of Si3 N4 (Ube, SN-ESP) with 6% Y2 O3 (H.C. Starck, grade C) and 4% Al2 O3 (Alcoa, A16) as sintering additives. The material containing only this composition was used as a reference and it is henceforth denoted as SN. To this matrix several different types of graphene were admixed as reinforcement materials. The rst one was an experimental multilayer graphene (MLG) nanosheets prepared by the mechanical milling method in MFA, Budapest.24 In their work it was demonstrated that the high efcient attrition milling (Union Process, 01HD-HDDM) is a successful tool for producing graphene multilayers. The results showed that intensive milling in ethanol proved to be a better way how to separate the graphite layers than in water or in air atmosphere. The 10 h milling resulted in the largest distance between separated adhering multilayers (can exceed 1 m) showing the highest degree of exfoliation and the graphene multilayers were composed of approximately 40 graphene layers on average. Beside that also three commercial products were used: two grades of exfoliated graphene nanoplatelets (xGnP-M-5 particle size 5 m, and xGnP-M-25 particle size 25 m)25 by XG Sciences and nanographene platelets (Angstron N006-010-P)26 by Angstron Materials. The powder mixture was milled in attritor mill in distilled water by using zirconia tank, zirconia media (1 mm in diameter) and zirconia agitator delta discs for 5 h, 4000 rpm. As a consequence, each material contained also some zirconia from the milling the amount of which was between 3.6 and 4.2 wt%, as it is shown in Table 1. In the case of graphene added powder mixtures, rst only the powder mixtures have been processed with the attrition milling at high rotation speed, 3000 rpm during 4.5 h. Different types of multilayer MLG or few-layer graphene FLG were used as reinforcement material for ceramic matrices. Thus in the next steps of milling, 1 or 3 wt% of graphene nanoplatelets were mixed with the silicon nitride-based powder mixture in attrition mill with low rotational speed, 600 rpm until 30 min. The mixtures had been dried and sieved with a lter with mesh size of 150 m. Green samples were obtained by dry pressing at 220 MPa. Samples prepared for HIP were oxidized at 400 C to eliminate the PEG. Hot isostatic pressing (HIP) was performed at 1700 C in high purity nitrogen by a two-step

Graphene type
a

ZrO2 4.2 3.6 3.6 3.9 3.6 3.6 3.6 3.6 3.6

0 1 1 1 1 3 3 3 3

Exfoliated graphene nanoplatelets xGnP-M-5 Exfoliated graphene nanoplatelets xGnP-M-25 Nano graphene platelets Angstron N006-610-P Multilayer graphene Exfoliated graphene nanoplatelets xGnP-M-5 Exfoliated graphene nanoplatelets xGnP-M-25 Nano graphene platelets Angstron N006-610-P Multilayer graphene

From milling.

sinter-HIP method using BN embedding powder at 20 MPa, with 3 h holding time. The heating rate did not exceed 25 C/min. Composition and notation of all experimental materials used in this study is summarized in Table 1. Wear behavior of the experimental materials was studied in dry sliding in air. The surfaces were carefully prepared by polishing down to surface roughness below 0.05 m where possible. It has to be pointed out that for higher volume fractions (3%) of carbon additives this was not entirely possible, as the residual porosity prevented us to reach such a low roughness. The wear testing was carried out on the tribometer HTT by CSM Instruments in dry conditions at the room temperature (24 C, air humidity 40%) using the pin-on-disk technique, where the tribological partner was a highly polished (roughness Ra < 0.25 m) Si3 N4 ball with 6 mm diameter. The applied load was 5 N, the sliding speed 10 cm/s and the sliding distance was 300 m. The tangential forces during the test were measured and friction coefcients calculated. The worn surfaces were subsequently observed and the wear regimes, damage type and micromechanisms were identied. The material losses (volume of the wear tracks) due to wear were measured by a high precision confocal microscope PLu neox 3D Optical Proler, by SENSOFAR, and then specic wear rates (W) were calculated in terms of the volume loss (V) per distance (L) and applied load (Fp ) according to the standard ISO 2080827 : W= V (mm3 /mN) L Fp

At the same time, the worn cap on the counter body was measured by a light microscope, its volume and then its wear rate calculated, too. The microstructure and the wear tracks created at the surfaces of the investigated materials were observed using scanning electron microscopy (JEOL JSM 7000). Porosity and average pore size was found using image analysis of micrographs of polished ceramographic sections. At least 1 mm2 of surface was documented and then treated using ImageJ 1.47e software.

P. Hvizdo s et al. / Journal of the European Ceramic Society 33 (2013) 23592364 Table 2 Properties of experimental materials. Material SN SN-15 SN-125 SN-1A SN-1M SN-35 SN-325 SN-3A SN-3M Porosity, % <0.5 5.5 6.5 2.6 3.4 20.1 22.3 22.4 10.0 Average pore size, m 4.1 4.5 3.1 3.3 13.2 15.5 15.7 7.2 HV5, GPa 15.4 14.6 15.1 14.6 16.4 7.7 7.4 8.1 9.0 0.4 0.2 0.3 0.6 0.4 0.1 0.1 0.2 0.1 KIC , MPa m1/2 6.9 7.8 8.6 8.9 9.9 0.4 0.4 0.2 0.4 0.4
1.0

2361

0.8

coefficient of friction

0.6

0.4

0.2

3. Results and discussion Microstructures of the experimental materials were studied in detail previously in Refs. 28, 29. Kun et al.28 prepared and carried out basic characterization of the same silicon nitridegraphene systems. It was found that in all cases good mixing and homogeneous distribution of graphene additives was achieved28 and the graphene platelets were well embedded within the matrix. It was also observed that graphene platelets induced porosity in matrix and reduced the size of Si3 N4 grains in resulting composites. Dusza et al.29 performed a statistical analysis of the grain sizes and found that while in the monolithic Si3 N4 the matrix grains had typically 0.4 m, the average matrix grains in all composites had narrow size distribution with maximum at approximately 0.2 m. The results of X-ray investigations showed only -Si3 N4 diffraction peaks, suggesting that transformation was complete.28,29 The monolithic silicon nitride was fully dense. Carbon ller phases made densication more difcult which lead to porosity increasing with increasing volume fraction of carbon phases. Graphene does not tend to extensive clustering as it is typical for CNT16 , but occasional stacks of several graphene platelets were present.29 Dusza et al.29 studied morphology of these particular graphene types and found out that the milling prepared MLG had size around 12 m, majority of the Angstron platelets had similar size but their distribution was rather wider (typically from 1 to 6 m). The XG Sciences materials correspond to the values given by the producer, i.e. 5 m and 25 m. The pores present in the composites were always associated with graphene platelets. Values of porosity and average pore size found by image analysis of ceramographic sections are shown in Table 2. One can see that the pore sizes differ for different graphene types and generally larger graphene platelets are connected to larger pores. The tests showed that the friction behavior of all materials under the testing conditions was very stable. In all cases the friction force stabilized over a short run-up phase, shorter than 1 m. Then the friction remained more or less stable (depending on material) but in no case it exhibited any signicant change that would suggest some important change in character of wear damage mechanisms over the 300 m sliding distance. Fig. 1 shows the comparison of average COF for all experimental materials. It can be seen that for used amounts of

0.0

SN-1x25

SN-3x25

SN-1x5

SN-3x5

SN-1A

SN-3A

SN-1M

SN-1CNT

SN-3M
SN-3M

SN

Fig. 1. Coefcient of friction for all materials compared with values for Si3 N4 with CNT.16

graphene (1 and 3 wt%) the COF remained basically the same, within the scatter. No signicant lubrication effect was observed, on the contrary, if anything, the friction had a tendency to increase. This is similar to the result of the previous study of Si3 N4 CNT composites16 , values of which are also given in Fig. 1. There, it was found, that only when the amount of CNTs reached 5 wt% and 10 wt%, COF decreased by about 20% and 40%, respectively, in comparison with the reference Si3 N4 . Wear resistance results are summarized in Fig. 2 which shows the specic wear rate values for all materials as a function of graphene content together with analogous results found for Si3 N4 CNT.16 Here, one can observe that the graphene containing composites are generally more wear resistant than Si3 N4 CNT materials. In materials with 1 wt% of graphene the results are very similar to those of the reference Si3 N4 . They differ within experimental error. However, in materials with 3 wt%
1E-4

wear rate [mm /N.m]

1E-5

1E-6

SN

SN-1A

SN-3x5

SN-1x5

SN-1M

SN-1x25

SN-1CNT

SN-3x25

SN-3A

Fig. 2. Specic wear rates of all experimental materials together with values for Si3 N4 with CNTs.16

SN-3CNT

SN-3CNT

2362
1E-4

P. Hvizdo s et al. / Journal of the European Ceramic Society 33 (2013) 23592364

ball wear rate [mm /N.m]

1E-5

1E-6

SN

SN-1A

SN-1M

SN-3A

SN-1x5

SN-1x25

SN-3x5

Fig. 3. Specic wear rates of counterbodies (silicon nitride balls).

of graphene signicant differences exist showing that various graphene types behave differently. The lowest wear resistance had materials with MLG platelets with the wear rate close to that of Si3 N4 3%CNT. Other three graphene containing ceramics showed signicant improvement of wear resistance which meant that wear rates were about 60% lower than that for the reference Si3 N4 . The tendency suggests that larger graphene platelets coincide with better wear resistance. Fig. 3 shows the wear rate of the counterbodies (silicon nitride balls). It completes the information about the wear damage of the tribological pairs. Perhaps surprisingly, it shows general tendency to increase wear rates of the counterbodies with increasing amount of graphene in the paired material. It is in agreement with increase of the coefcient of friction and greater roughness of the composites which is connected to their higher porosity. The wear resistance can be related to other mechanical properties. According to Ref. 28 adding the carbon-based llers lowered the bending strength and elastic modulus. The multilayered graphene containing composite had the highest strength and modulus of elasticity among those composites. Vickers indentation was used to measure hardness and fracture toughness, similarly as in Ref. 18. Indentation study was used because the small specimen size prevented the use of the standard fracture toughness test. However, in the materials with 3% of graphene, the microstructures with higher porosity did not allow formation of usable the indentation crack system, so only the results for composites with lower amount of graphene are reported (Table 2). The results show that addition of 1 wt% of graphene did not reduce the hardness, even though it lead to slight increase in porosity. The fracture toughness, where measured, increased with addition of graphene. Kvetkov et al.18 identied several types of toughening mechanisms in graphene containing composites, from crack deection and slow down of the crack propagation at graphene platelets, to crack bridging accompanied by pull-out of graphene and dissipation of crack energy. This improvement of fracture toughness may be the reason for improved wear resistance of the composites. Addition

SN-3x25

SN-3M

Fig. 4. (a) Very faint wear track in the reference Si3 N4 . (b) Polished surface of the reference Si3 N4 .

of 3 wt% graphene reduced the hardness due to high porosity. SEM observations of the wear tracks revealed the wear damage micromechanisms that took place in the contact zone. Fig. 4 shows the wear track in the reference material which was very small and it had a smooth look. The materials with 1 wt% of carbon phases seem to contain more plastic deformation in the wear tracks (Fig. 5). In other sites, however, details do not differ very much and are still similar to the character of the original polished surface but with more porosity (Fig. 6). Materials with 3 wt% of graphene contain many pores within which we can observe graphene platelets. Yet, even inside the wear tracks these platelets usually remain in their sites (Fig. 7) and even if they are dislodged from the matrix, no graphite-like transfer lm is observed (Fig. 8). Its absence is difcult to prove conclusively but it can be inferred from the fact that COF of composites was never lower than that in the case of monolithic Si3 N4 . The results then suggest, that graphene platelets are bonded very strongly with the silicon nitride matrix grains and do not provide a rich source of graphite, that could lubricate the contact

P. Hvizdo s et al. / Journal of the European Ceramic Society 33 (2013) 23592364

2363

Fig. 5. Wear track in SN-15 (1 wt% graphene).

Fig. 8. SN-35 wear track. Wear debris on the surface, graphene platelets rmly attached to the matrix inside the pore.

surfaces, as in the case of carbon nanobers containing zirconia systems.11,16 4. Conclusions Series of silicon nitride based composites containing 1 and 3 wt% of four types of graphene platelets were prepared. The results allow to conclude the following: - Carbon ller phases made densication more difcult which lead to porosity increasing with increasing volume fraction of carbon phases. Graphene does not tend to extensive clustering as it is typical for CNT, only occasional stacks of several graphene platelets are present. - Different types of graphene inuenced the behavior of the composite in different way. - The graphene platelets were well integrated in the microstructure, in cases of materials with 1 wt% graphene measurably improved the fracture toughness. The highest measured fracture toughness had Si3 N4 with 1% multilayered graphene. - For used volume fractions (1% and 3%) of all types of carbon additives no decrease of coefcient of friction was observed. No signicant graphite transfer lm was found. - Carbon phases did not participate in lubricating process. However, introduction of graphene did lead to better wear resistance. The best results were found for the materials with 3% of three commercially available graphene grades, the highest wear resistance had Si3 N4 3% (xG) graphene whose wear rate was 60% lower than that of the reference silicon nitride. Acknowledgements This work was supported by the Slovak Grant Agency (project VEGA 2/0120/10), by projects NanoCEXmat II, ITMS No: 26220120035, and CeKSiM, ITMS No: 26220120056, and by Hungarian Scientic Research Grant OTKA 105355 and Hungarian-Korean Technical Cooperation Center Foundation.

Fig. 6. Detail of the inside of the wear track, SN-15.

Fig. 7. SN-35 pores in the polished surface, graphene akes are visible. They are well dispersed and well integrated in the microstructure.

2364

P. Hvizdo s et al. / Journal of the European Ceramic Society 33 (2013) 23592364 17. Pfeifer J, Sfrn G, Wber F, Zsigmond V, Koszor O, Arat P, et al. Tribology study of silicon nitride-based nanocomposites with carbon additions. Mater Sci Forum 2010;659:2358. 18. Kvetkov L, Duszov A, Hvizdo s P, Dusza J, Kun P, Balzsi C. Fracture toughness and toughening mechanisms in graphene platelet reinforced Si3 N4 composites. Scripta Mater 2012;66:7936. 19. Balzsi C. Silicon nitride composites with different nanocarbon additives. J Korean Ceram Soc 2012;49:35262. 20. Balzsi C, Fnyi B, Hegman N, Kovr Z, Wber F, Vrtesy Z, et al. Development of CNT/Si3 N4 composites with improved mechanical and electrical properties. Compos Part B: Eng 2006;37:41824. 21. Walker LS, Marotto VR, Raee MA, Koratkar N, Corral EL. Toughening in graphene ceramic composites. ACS Nano 2011;5:318290. 22. Ramirez C, Figueiredo FM, Miranzo P, Poza P, Osendi MI. Graphene nanoplatelet/silicon nitride composites with high electrical conductivity. Carbon 2012;50:360715. 23. Ramirez C, Garzon L, Miranzo P, Osendi MI, Ocal C. Electrical conductivity maps in graphene nanoplatelet/silicon nitride composites using conducting scanning force microscopy. Carbon 2011;49:387380. 24. Kun P, Wber F, Balzsi C. Preparation and examination of multilayer graphene nanosheets by exfoliation of graphite in high efcient attritor mill. Cent Eur J Chem 2011;9:4751. 25. http://www.xgsciences.com 26. http://angstronmaterials.com 27. ISO 20808:2004(E). Fine ceramics determination of friction and wear characteristics of monolithic ceramics by ball-on-disc method; 2004. 28. Kun P, Tapaszt O, Wber F, Balzsi C. Determination of structural and mechanical properties of multilayer graphene added silicon nitride-based composites. Ceram Int 2012;38:2116. 29. Dusza J, Morgiel J, Duszov A, Kvetkov L, Nosko M, Kun P, et al. Microstructure and fracture toughness of Si3 N4 + graphene platelet composites. J Eur Ceram Soc 2012;32:338997. Pavol Hvizdo s, Ph.D., MSc., has been working in the Structural Ceramics sice, Slovakia since 1988. Department of Institute of Materials Research, Ko In 1996 he received his Ph.D. degree in material sciences from Technical University, Ko sice, Slovakia. In 2000 he was awarded the Marie Curie Fellowship of European Commission to work in Queen Mary University of London, UK. Later he worked as a Ramon y Cajal Fellow at Polytechnic University of Catalonia, Barcelona, Spain. His area of scientic interest includes microstructure, mechanical and physical properties of composite structural ceramics. His experience comprises microstructural characterization and mechanical testing of ceramics. Recently he has been working in the eld of contact behaviour (tribology and indentation) of nanocomposites. He has published papers in several international proceedings and major material science journals.

References
1. Peigney A, Laurent Ch, Flahaut E, Rousset A. Carbon nanotubes in novel ceramic matrix nanocomposites. Ceram Int 2000;26:67783. 2. Ma RZ, Wu J, Wei BQ, Liang J, Wu DH. Processing and properties of carbon nanotubesnano-SiC ceramic. J Mater Sci 1998;33:524363. 3. Duszov A, Dusza J, Tom sek K, Morgiel J, Blugan G, Kuebler J. Zirconia/carbon nanober composite. Scripta Mater 2008;58:5203. 4. Duszov A, Dusza J, Tom sek K, Blugan G, Kuebler J. Microstructure and properties of carbon nanotube/zirconia composite. J Eur Ceram Soc 2008;28:10237. 5. Mazaheri M, Mari D, Hesabi ZR, Schaller R, Fantozzi G. Multi-walled carbon nanotube/nanostructured zirconia composites: outstanding mechanical properties in a wide range of temperature. Comp Sci Technol 2011;71:93945. 6. Sun J, Gao L, Iwasa M, Nakayama T, Niihara K. Failure investigation of carbon nanotube/3Y-TZP nanocomposites. Ceram Int 2005;31:11314. 7. Cha SI, Kim KT, Lee KH, Mo CB, Hong SH. Strengthening and toughening of carbon nanotube reinforced alumina nanocomposite fabricated by molecular level mixing process. Scripta Mater 2005;53:7937. 8. Balzsi C, Knya Z, Wber F, Bir LP, Arat P. Preparation and characterization of carbon nanotube reinforced silicon nitride composites. Mater Sci Eng C 2003;23:11337. 9. Yamamoto G, Hashida T, Adachi K, Takagi T. Tribological properties of single-walled carbon nanotube solids. J Nanosci Nanotechnol 2008;8:266570. 10. Suh MS, Chae YH, Kim SS. Friction and wear behavior of structural ceramics sliding against zirconia. Wear 2008;264:8006. 11. Hvizdo s P, Puch y V, Duszov A, Dusza J. Tribological behavior of carbon nanoberzirconia composite. Scripta Mater 2010;63:2547. 12. An JW, You DH, Lim DS. Tribological properties of hot-pressed aluminaCNT composites. Wear 2003;255:67781. 13. Yamamoto G, Omori M, Yokomizo K, Hashida T, Adachi K. Structural characterization and frictional properties of carbon nanotube/alumina composites prepared by precursor method. Mater Sci Eng B 2008;148: 2659. 14. Jones MI, Hyuga H, Hirao K, Yamauchi Y. Tribology of silicon based nitride ceramics. J Aust Ceram Soc 2012;48:2137. 15. Gonzalez-Julian J, Schneider J, Miranzo P, Osendi MI, Belmonte M. Enhanced tribological performance of silicon nitride-based materials by adding carbon nanotubes. J Am Ceram Soc 2011;94: 25428. 16. Hvizdo s P, Puch y V, Duszov A, Dusza J, Balzsi C. Tribological and electrical properties of ceramic matrix composites with carbon nanotubes. Ceram Int 2012;38:566976.