Structure and properties of Mo-containing diamond-like carbon films produced by ion source assisted cathodic arc ion-plating

Applied Surface Science 286 (2013) 109–114

Contents lists available at ScienceDirect

Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc

Structure and properties of Mo-containing diamond-like carbon films produced by ion source assisted cathodic arc ion-plating L.L. Wang a , R.Y. Wang b , S.J. Yan a , R. Zhang a , B. Yang b , Z.D. Zhang a , Z.H. Huang a , D.J. Fu a,∗ a Key Laboratory of Artificial Micro- and Nano-Materials of Ministry of Education and School of Physics and Technology, Wuhan University, Wuhan 430072 China b School of Power & Mechanical Engineering, Wuhan University, Wuhan 430072, China

a r t i c l e

i n f o

Article history: Received 29 March 2013 Received in revised form 24 July 2013 Accepted 6 September 2013 Available online 13 September 2013 Keywords: Mo-DLC films Ion source Hardness Multilayer structure

a b s t r a c t Ion source assisted cathodic arc ion-plating was used to synthesize molybdenum containing diamondlike carbon films. The element of molybdenum is uniformly distributed in our sample as analyzed by Rutherford backscattering spectroscopy. The surface morphology of the films was analyzed by scanning electron microscope and atomic force microscope. The structure and bond state of the molybdenum containing diamond-like carbon films were characterized by X-ray diffraction, high resolution transmission electron microscopy, Raman spectra, and X-ray photoelectron spectroscopy. The Mo content in the films was controlled by varying of the acetylene gas flow rates. The root-mean square roughness of the asdeposited sample was found in the range of 1.5 nm. The hardness of 35 GPa has been achieved at the optimum conditions of synthesis. This can be attributed to formation multilayer structure during deposition process and the formation of hard molybdenum carbide phase with C=Mo bonding. The results show that ion source assisted cathodic arc ion-plating is an effective technique to fabricate metal-containing carbon films with controlled metal contents. © 2013 Elsevier B.V. All rights reserved.

1. Introduction A great attention was devoted to the diamond-like-carbon (DLC) films due to their outstanding properties, such as high mechanical strength and hardness, excellent chemical inertness, and exceptional friction and wear performance [1]. Its application in electronic chips, optical tools and devices, biology and medicine as protective coatings of sensors were studied by introducing new variables during the synthesis process [2]. DLC coatings also have applications in orthopedics, cardiovascular components, and waveguide because they are very hard, have low friction, are fully biocompatible and prevent leaching of metallic ions into the body [3]. In recent years there has been much interest in metalcontaining carbon (Me-C:H) films. Their mechanical, tribological, magnetic, electrical, and optical properties have been widely studied. It is known that these properties are closely related to their microstructures [4]. Among them the transition metal carbides TiC, WC and NbC are interesting group of nanocrystalline materials. The conventional coarse-grained materials of the transition metal carbides show high hardness, high melting points, and chemical inertness making them useful as wear-resistant materials. In addition, several transition metal carbides such as WC and MoC also

∗ Corresponding author. E-mail address: [email protected] (D.J. Fu). 0169-4332/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.apsusc.2013.09.029

have a catalytic activity comparable to the platinum metals suggesting potential use in many catalytic processes [5]. Addition of transition metal atoms in the carbon network can enhance adhesion between the DLC films and the substrates, and can reduce the intrinsic compressive stress of the films by forming metallic crystalline domains and metal-carbon nanocomposites [2]. DLC coatings were deposited by many methods using variety carbonaceous precursor materials. For example, ion implantation [6], pulsed laser deposition [7], filtered cathodic arc deposition, magnetron sputtering [8], and RF plasma activated chemical vapor deposition [3]. A drawback to pulsed laser deposition (PLD) is that occasionally droplets or chunks of the target material are ablated and cause a surface roughness that is not easily removed [3]. The disadvantages for ion implantation are crystalline damage of the substrate, low productivity and high expenditure of the equipment. High deposition temperature and low growth rate are the main drawback to chemical vapor deposition. Sputtering and cathode arc are preferred by industrial produce. However, sputtering always shows drawbacks of low-usage of the targets, instability of the plasma and low film growth rate, so now industrial produce more and more trend to cathode arc due to high ionization ratio, high plasma intense and fast growth rate [9]. Acetylene, methane, and graphite target were usually used as the carbon source of the DLC film in cathode arc deposition process. Acetylene is the preferred gas for low pressure deposition, because its strong C≡C bond means it has a simple dissociation pattern [9].

110

L.L. Wang et al. / Applied Surface Science 286 (2013) 109–114

17500

Yield (count/ch.)

15000

Experimental Simulated

Mo

12500

Si 10000

C

7500 5000 2500 0 100

150

200

250

300

350

Channel number Fig. 2. Proton beam backscattering spectrum of Mo-DLC film at C2 H2 flow rate 100 sccm on Si substrate with the thin Mo film as transition layer.

Fig. 1. Schematic diagram of the ion source assisted cathodic arc ion-plating system.

In this work DLC films containing transition metal Mo were prepared by ion source assisted cathodic arc ion-plating using acetylene gas. The aim of this article is to show the possibility of increasing [1,10] of hardness by formation of the multilayered DLC structure and investigate its microstructure and mechanical properties. 2. Experiment details 2.1. System description The top view schematic diagram of the ion source assist cathodic arc ion-plating system is shown in Fig. 1. The acetylene gas as the carbon source was introduced into the chamber through the hollow cathode and was ionized by the plasma. The hollow cathode ion source used in this work is to ensure the full ionization of the acetylene gas and the restrictions of the plasma to offer high dense ions on the other hand. The holder is used for fixing the substrates and it can rotate to ensure the uniform growth of the films. 2.2. Film deposition Mo containing DLC films were deposited on silicon (1 0 0) wafers, cemented carbides, and stainless steel substrates using ion source assisted cathodic arc ion-plating system with a Mo target and a hollow cathode ion source. The carbon source was offered by the hollow cathode through ionizing acetylene gas. The substrates were cleaned sequentially in methanol, ethanol, and deionized water, in each for 5 min at 60 ◦ C, and then dried with nitrogen gas. The substrates were placed vertically on a rotating sample holder in the center of the chamber. The pressure of the chamber was better than 7 × 10−3 Pa and the temperature above 300 ◦ C before deposition. A high bias voltage of −1000 V was applied to the substrate and the Mo target current set at 80 A to offer high energy ions, the substrate were further cleaned by this ions at the beginning of the deposition for 10 min. A bias voltage of −100 V was added to the substrates with 80% duty cycle and the chamber pressure was kept at 1.2 Pa during the film deposition process. A transition layer of Mo lasted 5 min was deposited onto the substrates to enhance the adhesion of the Mo-DLC films with substrates. The films were deposited 60 min under fixed current of 70 A of the Mo target and 50 A of the hollow cathode with different acetylene flow

rate (0, 50, 100, 150, 200 sccm). Argon gas was introduced in order to maintain the pressure of 1.2 Pa during deposition process. The temperature was kept at around 300 ◦ C during DLC films deposition process and decreased to room temperature after deposition. The distance between substrates and targets is about 10 cm. 2.3. Characterization The Rutherford backscattering spectroscopy (RBS) technique was carried out using a 1.74 MeV proton beam to determine the Mo and C contents as well as the film thickness. The scattered protons were detected at a backscattering angle of 170◦ by a silicon surface barrier detector with an energy resolution of ∼15 keV and an effective detection area of 50 mm2 . The distance between the detector and the sample is 112.6 mm and the solid angle ˝ is 3.94 msr. Spectral analysis was carried out using the iterative analytical simulation code SIMNRA. The surface morphology of the films were analyzed by field emission scanning electron microscope (FESEM) with EDAX genesis 7000 EDS system and SPM-9500J3 atomic force microscope (AFM) manufactured by SHIMADZU. The crystallinity and microstructure were analyzed by X-ray diffraction (XRD) performed on D8 advanced X-ray diffractometer with a Cu K␣ radiation (0.154 nm) and JEM-2010FEF (HT) transmission electron microscope operated at 200 kV. An RM-100 confocal Raman Micro-spectrometer with Ar+ laser excitation (632.8 nm) was used to determine the content of sp3 hybrid bond. The electron binding energy in C1s and Mo3d orbits was detected by XSAM800 X-ray photoelectron spectroscopy using Mg K␣ excitation (1253.6 eV). The microhardness were measured by using a HX-1000 Vickers microhardness tester on a load of 25 g. Ten random indentation measurements were conducted and the mean value was calculated for each sample. 3. Results and discussions 3.1. Morphology and structure In order to determine the contents and profiles of Mo and C in the films, RBS analysis was performed and the spectrum was fitted [11] using the SIMNRA program as shown in Fig. 2. It can be observed that molybdenum is uniformly distributed in our sample. According to the simulated result, the thickness of the DLC film is about 1.98 ␮m and the transition layer of Mo film is 320 nm. The experimental value of total thickness obtained from section SEM image is 2.23 ␮m. The deviation between simulated and measured value is 3.14%. The carbon content of the as grown films increases

L.L. Wang et al. / Applied Surface Science 286 (2013) 109–114

111

Fig. 3. Large area SEM images of the Mo-DLC film (a) and (b) the carbon content as a function of C2 H2 flow rate.

with increasing C2 H2 flow rate from 0 to 200 sccm, as shown in Fig. 3b, and reaches 75.0 at.% when the C2 H2 flow rate is 200 sccm. The large area top morphology of the as-deposited Mo containing DLC film is shown in Fig. 3a. A smooth surface with several particles is seen from this picture. This can attribute to low bias voltage of −100 V during DLC films deposition process, because large particles can’t get enough energy to reach the substrates at such low bias voltage. The surface images of the films obtained by AFM are shown in Fig. 4a and b. The relatively rougher morphology can be seen on the surface of the as-deposited films without C2 H2 addition, as shown in Fig. 4a. At presence of C2 H2 as shown in Fig. 4b it reveals that there are no large Mo clusters in the film and Mo macroparticles (with a size in the range of micrometers) generated from

the cathodic arc source have been effectively removed [11]. Root mean square roughness was calculated from AFM images and the results are shown in Fig. 4c. The C2 H2 absent specimen have showed a relatively higher roughness with a value of 2.249 nm, and the RMS∼1.5 nm for the coating deposited by using C2 H2 doped specimens. The roughness measurements indicate that Mo containing DLC films exhibit a smooth surface with the RMS around 1.5 nm. As can be seen in Fig. 5 most of the films consist of molybdenumcarbide and exhibit broad X-ray diffraction peaks in general. The line-broadening can be assigned to several factors such as a small grain size and stresses in the film [5]. The peak intensity of both crystalline phase Mo0.42 C0.58 and Mo2 C increases and then decreases with the C2 H2 flow rate increase from 50 to 200 sccm, and reach a maximum value when the C2 H2 flow rate is 100 sccm. The

Fig. 4. Surface morphologies of (a) Mo film and (b) Mo-DLC film and (c) root-mean-square roughness as a function of C2 H2 flow rate.

112

L.L. Wang et al. / Applied Surface Science 286 (2013) 109–114

Mo2C(310)

Intensity (a.u.)

Mo0.42C0.58(111)

C2H2 200 sccm 150 sccm 100 sccm 50 sccm 0 sccm

Mo (110)

20

30

40

50

60

70

2-Theta (deg.) Fig. 5. X-Ray diffraction of Mo-DLC films deposited at different C2 H2 flow rates.

pure metal Mo film was synthesized for comparison. The interplanar lattice spacing calculated using the diffraction law (2d sin=) is 0.224 and 0.153 nm, corresponding to phase Mo0.42 C0.58 and Mo2 C, respectively. These interplanar spacing values agree with the values of 0.211 and 0.153 nm obtained from HRTEM image in Fig. 6b. Fig. 6a shows a high resolution TEM image of specimen with the C2 H2 flow rate of 100 sccm. A multilayer structure was formed by itself due to the rotation of the substrate holder during the deposition process. This self-formed multilayer structure can release the inner stress [12] of the DLC film and improve the hardness (this is confirmed by microhardness measurement). A small grain size of about 3–5 nm is observed in the higher magnification image of this sample shown in Fig. 6b. The common structure consisting of crystalline grains with an amorphous interlayer in thin film structures is observed. From simple calculation, the interplanar spacing of the crystalline grains is 0.21 and 0.1408 nm, consistent with the values of 0.224 and 0.153 nm calculated from XRD. The effective way to distinguish the microstructures of DLC films is Raman spectroscopy. Fig. 7a shows the Raman spectra of Mo-DLC films with various metal concentrations. Three peaks related to the formation of molybdenum-carbide in the range of 500–1000 cm−1 are observed for all samples [13,14]. The intensities of D peak around 1350 cm−1 and G peak around 1550 cm−1 are enhanced

gradually with the C2 H2 flow rate increase from 50 to 200 sccm. It indicates more and more carbon host matrixes are formed. This is because of the G peak is due to the E2g C–C stretching vibration, and the D peak is due to disordering of ordered graphite structure [9,15]. The profile-fitting procedure is needed to derive various parameters for evaluating the carbon host matrix [16]. The ID to IG ratio is one of the parameters and its tendency as a function of C2 H2 flux is shown in Fig. 7b. The ratio increases with the C2 H2 flux, it suggests that the films become more graphite-like [11]. Many previous studies have shown that the G peak shifts towards higher wavelength and the ratio of ID /IG increases as the sp3 fraction decreases [9]. The G peak shifts towards higher wavelength and the ID /IG ratio increases as the C2 H2 flux increase in our work. The great decrease in the internal stress also indicates the decrease in sp3 hybridization [1]. The C1s XPS spectra fitted by Gaussian function fitting method for different C2 H2 flow rate are shown in Fig. 8a–d. The C 1s peak is consisted by four different carbon binding peaks: C=Mo double bond, carbon sp2 hybridization, carbon sp3 hybridization and C=O double bond which located at 283.4, 284.9, 285.8 and 288.7 eV, respectively [1,4,11,17]. The content of sp3 decreases and then increases, and reaches a minimum of 60% at the 100 sccm of the C2 H2 flow rate. This indicates a low compressive stress of the films, because sp3 sites are the source of compressive stress and good molybdenum-carbide phases are formed at this C2 H2 flux [1]. The hardness of all the Mo-DLC films is between 28 and 35 GPa. The high hardness can be attributed to the content of sp3 , which is 60–85%. The sp3 bonding is responsible for many properties of DLC including high hardness and chemical inertness [9]. Namely, the molybdenum-carbide phases would be beneficial to generating a high density of interfaces and a high volume of grain boundaries so as to relax the stress by the grain boundary diffusion and grain boundary sliding, leading to a lowered internal stress [17,18]. In contrary, the content of C=Mo double bond increases and then decreases, and reaches a maximum of 10% at the 100 sccm of the C2 H2 flow rate. This high content of C=Mo corresponds to the maximum hardness of the films, attributed to the formation of molybdenum carbides in these films that can improve the hardness of DLC films [19]. The content tendency of carbon sp3 hybridization bond and C=Mo double bond are summarized in Fig. 8f. The C=O double bond detected in the films is attributed mainly to air exposure [20]. Fig. 8e shows the Mo3d peaks at different C2 H2 flow rate. It is observed the Mo3d5/2 and Mo3d3/2 peaks become broader

Fig. 6. High resolution TEM images of Mo-DLC films deposited at the C2 H2 flow rate 100 sccm (a) and higher magnification image of this sample (b).

L.L. Wang et al. / Applied Surface Science 286 (2013) 109–114

3000

(a)

MoC

D

G

2500

2.0

200 150

2000 100 50

1500

(b)

1.5

ID/IG

Intensity (a.u.)

C2H2 flow rate(sccm)

1.0

0.5

100

500

1000

1500 -1

2000

113

120

140

160

180

200

C2H2 flow rate (sccm)

Raman shift (cm ) Fig. 7. Raman spectra and ID /IG ratio of the films deposited at different C2 H2 flow rates.

Fig. 8. Typical XPS of Mo-DLC films: C1s XPS fitted by using Gaussian function for different C2 H2 flow rates: (a) 50 sccm, (b) 100 sccm, (c) 150 sccm, and (d) 200 sccm. (e) Mo3d XPS spectra at different C2 H2 flow rates. (f) C=Mo content in the C1s XPS spectra as a function of C2 H2 flow rate.

114

L.L. Wang et al. / Applied Surface Science 286 (2013) 109–114

pure Mo metal (13 GPa). It is also much higher than that of the MoDLC film (10–11 [1], 24 [18], 23 GPa [10]) obtained in other previous studies. The improved hardness of Mo-DLC films can be attributed to the self-formed multilayer structure as shown in Fig. 6a and high content of the molybdenum-carbide phase as displayed in Fig. 8f.

80

AE(a.u.)

60

4. Conclusion

40

20

0 0

20

40

60

80

100

Load (N) Fig. 9. Scratch test curve of the Mo-DLC film prepared at the C2 H2 flow rate of 100 sccm.

35

Hardness (Gpa)

30

Mo-DLC films have been synthesized using ion source assisted cathodic arc ion-plating system. The experimental results show that Mo is uniformly distributed in the film and the Mo concentration can be controlled by varying the C2 H2 flow rates. The films exhibit extremely smooth surface with RMS∼1.5 nm and high hardness of 35 GPa at the optimum condition. XPS and XRD results show the formation of metallic carbide inside the amorphous carbon matrix. High resolution TEM results show the multilayer structure formed automatically during deposition process. The promotion of C=Mo bond content also has been displayed by fitting the C1s XPS spectra. All of these crystal structures and bonding states are related to the improved hardness of the Mo-DLC. Further work to determine the other properties of the films such as mechanical, optical, conductive properties will be done in the future. Acknowledgment

25

This work was supported by National Natural Science Foundation of China under 11275141 and 11105100, International Cooperation Program of Ministry of Science and Technology of China under 2011DFR50580, and by Fundamental Research Funds for Central Universities under 201120202020004.

20 15

References 10

0

50

100

150

200

C2H2 flow rate (sccm) Fig. 10. Hardness of Mo-DLC films as a function of C2 H2 flow rate.

and the intensity of these two peaks decrease with the increase of the C2 H2 flow rate. The peaks of Mo3d5/2 and Mo3d3/2 located at 227.9 and 231.0 eV belong to Mo–Mo binding. Both of these peaks exhibit shift to high binding energy edge of the Mo–carbide binding, and the shift increases (228.4–228.7 eV) for Mo3d5/2 and 231.6 to 231.9 eV for Mo3d3/2 with the C2 H2 flow rate increases [1,11]. 3.2. Mechanical property Fig. 9 shows a curve of scratch test for the Mo-DLC film prepared at a C2 H2 flow rate of 100 sccm. Acoustic emission were recorded during the test versus normal load (L) and a sharp rising on the curve implies that failure occurs, which can be considered as a criterion for determination of the critical load (Lc ) [21]. As shown in Fig. 9, the critical load is 70 N. The good adhesion result from the fact that selfformed multilayer structure and formation of Mo–carbide phases release the internal stress of the films. Fig. 10 shows the hardness of the Mo-DLC films deposited at different C2 H2 flow rate. It is seen that the DLC films containing Mo have a hardness of 28–35 GPa, which are higher than that of the

[1] L. Ji, H.X. Li, F. Zhao, W.L. Quan, J.M. Chen, H.D. Zhou, Appl. Surf. Sci. 255 (2009) 4180. [2] C. Corbella, E. Bertran, M.C. Polo, E. Pascual, J.L. Andújar, Diamond Relat. Mater. 16 (2007) 1828. [3] G. Dearnaley, J.H. Arps, Surf. Coat. Technol. 200 (2005) 2518. [4] Q.F. Huang, S.F. Yoon, Rusli, K. Chew, J. Ahn, J. Appl. Phys. 90 (2001) 4520. [5] J. Lu, U. Jansson, Thin Solid Films 396 (2001) 53. [6] S. Anders, A. Anders, I.E. Brown, B. Wei, J.W. Ager, K.M. Yu, Surf. Coat. Technol. 388 (1994) 68. [7] A.A. Voevodin, M.S. Donley, J.S. Zabinski, Surf. Coat. Technol. 92 (1997) 42. [8] X.L. Peng, Z.H. Barber, T.W. Clyne, Surf. Coat. Technol. 138 (2001) 23. [9] J. Robertson, Mater. Sci. Eng. R37 (2002) 129. ˇ ´ A. Jäger, Surf. Coat. Technol. 205 (2010) 1486. [10] P. Novák, J. Musil, R. Cerstv y, [11] R.K.Y. Fu, Y.F. Mei, L.R. Shen, G.G. Siu, K. Chu Paul, W.Y. Cheung, S.P. Wong, Surf. Coat. Technol. 186 (2004) 112. [12] B. Yang, R.Y. Wang, D.J. Fu, H. Ding, Chinese Patent No. CN102864411A. [13] B. Yang, Z.H. Huang, H.T. Gao, X.J. Fan, D.J. Fu, Surf. Coat. Technol. 201 (2007) 6808. [14] D. Lin-Vien, N.B. Colthurp, W.G. Fateley, J.G. Grasselli, Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules, Academic, New York, 1991. [15] W.J. Yang, Y.H. Choa, T. Sekino, K.B. Shim, K. Niihara, K.H. Auh, Thin Solid Films 434 (2003) 49. [16] T. Takeno, Y. Hoshi, H. Miki, T. Takagi, Diamond Relat. Mater. 17 (2008) 1669. [17] S. Zhang, D. Sun, Y.Q. Fu, H.J. Du, Surf. Coat. Technol. 198 (2005) 2. [18] J. Schiotz, F.D.D. Tolla, K.W. Jacobsen, Nature 391 (1998) 561. [19] S.F. Rusli, Q.F. Yoon, H. Huang, M.B. Yang, J. Yu, Q. Ahn, Zhang, J. Appl. Phys. 88 (2000) 3699. [20] Q.F. Huang, S.F. Yoon, Rusli, H. Yang, J. Ahn, Q. Zhang, Diamond Relat. Mater. 9 (2000) 534. [21] J.Z. Tian, Q. Zhang, Q. Zhou, S.F. Yoon, J. Ahn, S.G. Wang, J.Q. Li, D.J. Yang, Appl. Surf. Sci. 239 (2005) 255.