- Email: [email protected]
Structural evolution of Ti destroyable interlayer in large-size diamond film deposition by DC arc plasma jet
Accepted Manuscript Title: Structural evolution of Ti destroyable interlayer in large-size diamond film deposition by DC arc plasma jet Author: Jianchao Guo Chengming Li Jinlong Liu Junjun Wei Liangxian Chen Chenyi Hua Xiongbo Yan PII: DOI: Reference:
S0169-4332(16)30336-1 http://dx.doi.org/doi:10.1016/j.apsusc.2016.02.158 APSUSC 32676
To appear in:
APSUSC
Received date: Accepted date:
2-12-2015 18-2-2016
Please cite this article as: Jianchao Guo, Chengming Li, Jinlong Liu, Junjun Wei, Liangxian Chen, Chenyi Hua, Xiongbo Yan, Structural evolution of Ti destroyable interlayer in large-size diamond film deposition by DC arc plasma jet, Applied Surface Science http://dx.doi.org/10.1016/j.apsusc.2016.02.158 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Structural evolution of Ti destroyable interlayer in large-size diamond film deposition by DC arc plasma jet Jianchao Guo, Chengming Li, Jinlong Liu, Junjun Wei, Liangxian Chen, Chenyi Hua, Xiongbo Yan Institute of Advanced Materials and Technology, University of Science and Technology Beijing, Beijing 100083, P.R. China Graphical abstract
Highlights 1 The evolution processes of a Ti modeled interlayer were evaluated in DC arc plasma. 2 Sandwich structures were observed along with the growth of TiC. 3 Ti interlayer released stress via self-fracture.
Abstract: The addition of titanium (Ti) interlayer was verified to reduce the residual stress of diamond films by self-fracturing and facilitate the harvest of a crack-free free-standing diamond film prepared by direct current (DC) arc plasma jet. In this study, the evolution of the Ti interlayer between large-area diamond film and substrate was studied and modeled in detail.
Corresponding author. E-mail address: [email protected] (C. M. Li)
The evolution of the interlayer was found to be relevant to the distribution of the DC arc plasma, which can be divided into three areas (arc center, arc main, and arc edge). The formation rate of titanium carbide (TiC) in the arc main was faster than in the other two areas and resulted in the preferred generation of crack in the diamond film in the arc main during cooling. Sandwich structures were formed along with the growth of TiC until the complete transformation of the Ti interlayer. The interlayer released stress via self-fracture. Avoiding uneven fragile regions that formed locally in the interlayer and achieving cooperatively released stress are crucial for the preparation of large crack-free diamond films. Keywords: free-standing diamond film; DC arc plasma jet; Ti destroyable interlayer; evolution process
1. Introduction Polycrystalline diamond film has several exceptional properties that make it an outstanding optical material or other functional materials [1–3]. The demand for diamond films has increased in many applications, especially for large-area diamond films [4,5]. Among all deposition methods for large-area polycrystalline diamond film, direct current (DC) arc plasma jet chemical vapor deposition (CVD) is one of the most popular techniques because it can prepare large-area free-standing diamond films with a high deposition rate [6–8]. In the DC arc plasma jet CVD, the high-energy activation process by DC voltage leads to increased plasma temperature for C–H or C–H–O radicals. A higher deposition rate can be reached compared with that through microwave CVD or other methods. Nevertheless, synthesizing a large freestanding diamond film without crack remains a challenge because of the different heat expansion coefficients between prepared diamond films and the substrate. Large residual stress in diamond films is undoubtedly the primary cause of diamond film fractures [9]. The low success ratio of crack-free diamond films restricts the development and application of largearea diamond films. The introduction of an interlayer is an effective method to reduce the residual stress of diamond films [10–12]. Among all conventional interlayer materials, titanium (Ti) is the primary element employed extensively for diamond film preparation. Ti can release the stress of a diamond film via self-fracture, and hence improve the success rate of a crack-free diamond film [9,13,14]. Prior to diamond deposition, the titanium carbide (TiC) layer is formed first, which is important for diamond nucleation and growth [15]. In particular, the conversion degree from Ti interlayer to TiC is affected by the deposition environment. For example, the TiC layer
can be as thick as 150 μm or too thin to be detected under different conditions [16,17]. As for the DC arc jet CVD method, the DC arc plasma distributions above the substrate surface can be divided into three areas, namely, arc center, arc main, and arc edge [18]. These areas have an impact on TiC formation. In the present paper, polycrystalline diamond films (120 mm in diameter) were deposited on the graphite substrate with a Ti interlayer by DC arc plasma jet CVD operating at the gas recycling mode. The correlations among the evolution process of Ti interlayer, distributions of residual stress, and typical fracture mode in large-size diamond films were analyzed and discussed by detecting the variation of microstructures and compositions of Ti interlayer in different divided areas with various deposition times. 2. Experimental Polycrystalline diamond films (120 mm in diameter) were deposited on the Ti/graphite substrate via DC arc plasma jet. The equipment and processes had been discussed previously [9,19]. The Ti interlayer (approximately 8 μm thick) was deposited using multi-arc ion plating equipment. The diamond films were deposited for different deposition times (i.e., 1, 26, 72, and 152 h). The process parameters used in this study were as follows: power, 28–31 KW; substrate temperature, 800–950 ℃; chamber pressure, 2.5–3.8 KPa; CH4 concentration, 1 %. Apart from that in the nucleation stage, the deposition rate of diamond films was about 10 μm/h. The substrate surface was divided into three areas according to arc characteristics. After deposition of the diamond film, the graphite layers of samples were gently ground to a thickness of 0.1 mm to further confirm the correlation between the typical fracture mode of large diamond films and the unsynchronized evolution of the Ti interlayer caused by arc
characteristics. Scanning electron microscopy (SEM) of the QUANTAFEG 250 system was used to examine the morphology of the interlayer cross section. The distributions of C and Ti across the interface were evaluated through backscattered electron mode. Changes in Ti and C concentrations on the cross section of samples were evaluated using the line scan pattern energy dispersive spectrometer (EDS) available in the SEM system. The SMARTLAB X-ray diffraction (XRD) equipment was used to confirm the presence of TiC in the Ti interlayer. Furthermore, to clearly show the effect of the Ti interlayer on stress, the XRD sin2ψ method was used to statistically determine the residual stress in CVD diamond films deposited on graphite substrates with a Ti interlayer. The tests were conducted from the nucleation surface of the diamond because the diamond films became thick after deposition for hundreds of hours, and a discrepancy in growth and nucleation sides of the diamond films was found [20]. 3. Results and discussion 3.1. Arc characteristic of DC arc plasma jet The DC arc plasma generated between cathode and circle anode has a point-to-point discharge pattern. The arc plasma driven by external magnetic fields and rotated gas is in highspeed rotating mode, which can reach several thousand rpm. Given that the arc plasma is affected by electric, magnetic, and flow fields, the reaction is extremely complicated. In this study, the static image of DC arc plasma was obtained using a high-speed camera (Fig. 1). On the basis of the shape characteristics, arc distributions above the surface space of a substrate can be divided into arc center, arc main, and arc edge [18]. Correspondingly, three areas on the substrate surface were labeled as area 1, area 2, and area 3. The arc main is closer to the
substrate compared with the other two regions. The special shape of the arc situated the Ti interlayer on the substrate in a non-homogeneous environment, which caused different evolution processes for Ti interlayer in different areas. 3.2. Evolution process of the Ti interlayer 3.2.1. Changes in sectional morphology characteristics and element concentration distributions Fig. 2 shows the columnar grains in the Ti interlayer before the deposition of the diamond film. The sectional morphology characteristics and element concentration distributions of the Ti interlayer in three divided areas at four deposition times are shown in Fig. 3. The fracture morphology from the top of the diamond, Ti transition layer, and graphite substrate is also displayed. Figs. 3(a1)–(a3) show that the Ti interlayer began to gradually transfer to the isometric grain, and the columnar crystal nearly disappeared after 1 h of diamond film deposition. During the first 26 h, the interlayer thickened further and fine-grain layers appeared on the interfaces of the Ti interlayer. The sandwich structure was formed, and no significant differences among the three different areas were found, as shown in Figs. 3(b1)–(b3). The fine grains met in the center of the Ti layer within 72 h in areas 1 and 2, as shown in Figs. 3(c1) and 3(c2) respectively. The grain size of the Ti interlayer in area 2 was clearly smaller than that in area 1 from 72 h to 152 h, which proved that the extending rate in area 1 was slower than in area 2. The growth rate of the fine-grain layer in area 3 was significantly lower than that in the other two areas with prolonged deposition times. From Figs. 3(d2)–(d3), numerous cracks were found in the fine-grain-enriched regions when
deposition time was over 152 h, especially in area 2. Thus, the release of stress destroyed the interlayer. The cracks could be divided into horizontal and vertical directions according to the propagation orientation of cracks. Horizontal cracks appeared on the center of the interlayer (the intersectional region of small grains from two sides) and the junction regions of Ti layer and two carbon layers (diamond layer and graphite layer). The positions of the vertical cracks were stochastic. Area 2 is taken as an example to determine the relationship between fine grains and element distributions with different deposition times. The element distributions in the cross–section depth of samples were evaluated by line scanning of the EDS spectrum. The results are shown in Figs. 3(a2)–(d2). At the preliminary stage of diamond film deposition for 1 h [Fig. 3(a2)], no obvious diffusion of Ti and C could be found in the interlayer. Fig. 3(b2) shows that when deposition times increased to 26 h, fine-grain layers appeared on both sides of the Ti layer. Correspondingly, a dramatic diffusion of Ti and C proceeded in both interfaces. As deposition time increased to 72 h, the C element clearly distributed in the interlayer, and the carbon content was relatively high in small-grain regions, as shown in Fig. 3(c2). Figs. 3(c2) and 3(d2) show that only weak variations of the concentration distributions were found when the deposition times increased from 72 h to 152 h. Changes in C concentration were synchronized with the extension of fine grains shown in the SEM results. Thus, the fine grains were probably the carbonization products of the Ti interlayer. 3.2.2. Changes in X-ray diffraction spectra Fig. 4 shows the X-ray diffraction spectra of the samples coming from three divided areas of nucleation surface of diamond films deposited for different time periods. Clear diffraction
peaks of Ti, TiC, TiO, diamond, and graphite are shown in three divided areas for 1 h of deposition [Figs. 4(a1)–(a3)]. TiO should be generated during coating or chemical reaction with air before diamond film deposition. The TiC between diamond and Ti interlayer was formed because of the carbon atom in the deposition atmosphere defusing into the Ti layer and between graphite and the Ti interlayer from substrate. With increased deposition times, diffraction peaks of the diamond in every area were significantly enhanced, and TiC diffraction peaks were much more apparent in areas 1 and 2 [Figs. 4(b1)–(b3)]. Ti diffraction peaks still appeared in 26 h and 72 h of deposition, as shown in Figs. 4(b1)–(c3). These peaks corresponded to the sandwich structure in Fig. 3. TiC diffraction peaks were enhanced, and Ti diffraction peaks disappeared when deposition times increased from 72 h to 152 h in Figs. 4(d1)–(d3). This result demonstrated the completion of the carbonization process to some degree. 3.2.3. Model and discussion of the evolution process of Ti interlayer The evolution process of the Ti interlayer is schematically displayed in Fig. 5. Fig. 5(a) shows that the columnar grains in the Ti interlayer are exhibited before diamond film deposition. Fig. 5(b) shows that columnar crystals have a declining trend after nucleation, which may be caused by recrystallization at high temperature. The development of sandwich structures (Fig. 5c) causes TiC to come into contact with both sides of the interlayer. A dividing line can also be seen in Figs. 3(d1)–(d3). The crack at the completion of diamond film deposition may have originated from the interfaces of TiC/TiC in the interlayer, diamond/TiC, and graphite/TiC. The significance of the destroyable interlayer is demonstrated in Figs. 3(d1)– (d3) and Fig. 5(d). In addition, given the elasticity modulus of diamond, TiC, and graphite at 700–1200 GPa, 448–451 GPa, and 4.8–28 GPa, respectively, the stress buffer is formed during
cooling because of the gradient of elasticity modules. Under the increased stress conditions, the interlayer uses self-fracture to release stress during cooling after diamond deposition. 3.3. Residual stress distributions of diamond films The residual stress distributions in diamond films after cooling were inevitably influenced by the uneven evolution of the Ti interlayer. The XRD sin2ψ method was used to determine residual stress in CVD diamond films for different deposition time periods in three divided areas. The (311) plane of CVD diamond was used with a tilt angle (ψ) from 0° to 45°. To guarantee the reliability of detection results, three points in each divided area were detected. The statistical residual stress distributions of diamond films in divided areas for different deposition times are shown in Fig. 6. Given that the diamond grains had not yet formed a complete film, the samples deposited for 1 h were not considered. With the increase of deposition time, the thickness of diamond films after different deposition times were about 250, 710 and 1500 μm, respectively. There was no obvious thickness difference of diamond films in three different areas. The thermal expansion coefficient mismatch among the interlayer and diamond film caused large compressive stress states to be detected in all areas of the diamond film after different deposition times. Fig. 6(a) shows that, during 26 h of deposition, the magnitudes of residual stress in three areas showed no obvious difference and each standard deviation was relatively small. Stress distributions ranging from –1.8 GPa to –2.15 GPa followed a slightly increasing trend from area 1 to area 3. Fig. 6(b) shows that after deposition for 72 h, the average stress of diamond films was slightly increased and stress distribution trends remained unchanged. However, the standard deviations of stress results in area 2 were significantly higher than in
the other two areas. Fig. 6(c) shows that when deposition times reached 152 h, the residual stress in areas 2 and 3 were reduced. The average stress in area 2 was –1.55 GPa, which was the lowest among three divided areas. SEM results showed more cracks in area 2 than in the other two areas, which reduced stress and caused uneven distributions of stress in diamond films, and a large stress gradient between areas 1 and 2. In summary, with less than 72 h of deposition time, only a part of the Ti interlayer transformed into TiC on both sides of Ti/graphite and diamond/Ti. The Ti layer had a low elastic modulus, and thus elastic deformation of the interlayer would gradually increase the compressive stress from center to edge in diamond films after cooling. When the deposition time was long enough, almost all the Ti interlayer transformed into TiC and more stress-fragile regions in area 2 were found. Achieving elastic deformation is difficult because of the relatively high elastic modulus of TiC. Thus, the interlayer in area 2 was first destroyed during the cooling process and stress was drastically released. 3.4. Typical cracks in diamond films Fig. 7 presents the typical cracks in diamond films (120 mm in diameter) deposited for almost 150 h. Main ring cracks were found in area 2, and parts of the cracks extended to area 3. An analysis of the characteristics of typical cracks indicated that the fracture mode is likely related to Ti interlayer features caused by arc area characteristics. Diamond films in area 2 had the largest residual stress drop, and more cracks correspondingly appeared in the interlayer. Thus, stress-releasing behavior occurred in area 2 due to the fragile regions formed in the Ti interlayer. Theoretical simulation showed that the temperature in area 1 was the highest, which decreased with increased distance to area 1 [21]. The special “L”-type rotating arc and the
intake method of carbon gases caused the highest carbon concentration to be achieved in area 2. Relatively high temperature and the highest carbon concentration were observed near the substrate surface in area 2. The transformation rate was the fastest and resulted in more fragile regions appearing in the interlayer of area 2. Thus, area 2 was not conducive to cooperative stress release in different areas during cooling. As described above, the unsynchronized transformation of the Ti interlayer caused uneven distributions of residual stress and an increased fracture probability in diamond films, especially in area 2. Thus, cooperative stress release of interlayer in different areas was crucial for preparing large crack-free diamond films. 4. Conclusions The introduction of the Ti interlayer effectively reduces the residual stress of diamond films in high-power DC arc plasma jet CVD operating at the gas recycling mode. With increased deposition times, the TiC layer formed on both sides of the Ti interlayer and extended inward at different rates. Sandwich structures were formed along with TiC growth until the complete transformation of the Ti interlayer. The unsynchronized evolution of the Ti interlayer caused uneven distributions of residual stress and led to typical ring cracks in large-size diamond films.
Acknowledgments This work was sponsored by the National Natural Science Foundation of China (No.51272024) and the Ph.D. Programs Foundation of the Ministry of Education of China (No.20110006110011). The authors deeply appreciate their financial support.
References [1] S. Coe, R. Sussmann, RS, Optical, thermal and mechanical properties of CVD diamond, Diamond Relat. Mater. 9 (2000) 1726-1729. [2] R. Sussmann, J. Brandon, G. Scarsbrook, C. Sweeney, T. Valentine, A. Whitehead, C. Wort, Properties of bulk polycrystalline CVD diamond, Diamond Relat. Mater. 3 (1994) 303-312. [3] A. Gicquel, K. Hassouni, F. Silva, J. Achard, CVD diamond films: from growth to applications, Curr. Appl. Phys. 1 (2001) 479-496. [4] A. Kromka, O. Babchenko, T. Izak, K. Hruska, B. Rezek, Linear antenna microwave plasma CVD deposition of diamond films over large areas, Vacuum. 86 (2012) 776-779. [5] J. Weng, L. Xiong, J. Wang, S. Dai, W. Man, F. Liu, Investigation of depositing large area uniform diamond films in multi-mode MPCVD chamber, Diamond Relat. Mater. 30 (2012) 15-19. [6] J.V. Busch, J.P. Dismukes, Trends and market perspectives for CVD diamond, Diamond Relat Mater. 3 (1994) 295-302. [7] S. Matsumoto, Y. Manabe, Y. Hibino, Diamond deposition using anX-Y stage in a dc plasma jet chemical vapour deposition, J. Mater. Sci. 27 (1992) 5905-5910. [8] N. Ohtake, M. Yoshikawa, Nucleation effects and characteristics of diamond film grown by arc discharge plasma jet chemical vapor deposition, Thin solid films. 212 (1992) 112121. [9] C. Li, L. Wang, L. Chen, Z. Liu, L. Hei, F. Lu, Free-standing diamond films deposited by DC arc plasma jet on graphite substrates with a destroyable Ti interlayer, Diamond Relat. Mater. 18 (2009) 1348-1352. [10] R. Polini, F.P. Mantini, M. Barletta, R. Valle, F. Casadei, Hot filament chemical vapour deposition and wear resistance of diamond films on WC-Co substrates coated using PVDarc deposition technique, Diamond Relat. Mater. 15 (2006) 1284-1291. [11] X. Peng, T. Clyne, Formation and adhesion of hot filament CVD diamond films on titanium substrates, Thin Solid Films. 293 (1997) 261-269. [12] Y. Li, Y. Tang, Q. Yang, C. Xiao, A. Hirose, Diamond deposition on steel substrates with an Al interlayer, Int. J. Refract. Met. Hard Mater. 27 (2009) 417-420. [13] Y. Zhu, L. Wang, W. Yao, L. Cao, Interface diffusion and reaction between Ti layer and Si3N4/Si substrate, Surf. Interface Anal. 32 (2001) 296-300. [14] E. Buccioni, E. Braca, J. Kenny, M. Terranova, Processing-structure-adhesion relationship in CVD diamond films on titanium substrates, Diamond Relat. Mater. 8 (1999) 17-24. [15] T. Grogler, E. Zeiler, R. Singer, S. Rosiwal, M. Dannenfeldt, Growth stages of chemical vapor deposited diamond on the titanium alloy Ti-6Al-4V, Diamond Relat. Mater. 6 (1997) 1658-1667. [16] Q. Chen, Z. Lin, Diamond growth on thin Ti wafers via chemical vapor deposition. J, Mater. Res. 10 (1995) 2685-2688. [17] S.S. Park, J.Y. Lee, Synthesis of diamond films on titanium substrates by hot-filament chemical vapor deposition, Appl. Phys. 69 (1991) 2618-2622. [18] C. Li, R. Zhu, J. Liu, L. Chen, J. Guo, C. Hua, L. Hei, J. Wei, J. Wang, Z. Feng, Effect of
arc characteristics on the properties of large size diamond wafer prepared by DC arc plasma jet CVD, Diamond Relat. 39 (2013) 47-52. [19] F. Lu, G. Zhong, J.G. Sun, Y. Fu, W. Tang, J. Wang, G. Li, J. Zang, C. Pan, C. Tang, A new type of DC arc plasma torch for low cost large area diamond deposition, Diamond Relat. Mater. 7 (1998) 737-741. [20] R. Zhu, C. Li, L. Chen, J. Liu, J. Wei, L. Hei, H. Guo, Erosion difference of growth and nucleation sides of free-standing diamond films prepared by DC arc plasma jet CVD, Appl. Surf. Sci. 355 (2015) 203-208. [21] T. Huang, W. Tang, F. Lu, J. Gracio, N. Ali, Argon-to-hydrogen ratio in plasma jet diamond chemical vapour deposition, Surf. Coat. Technol. 190 (2005) 48-53.
Figure captions Fig. 1. Static state image of the arc and arc distributions of the DC arc plasma jet CVD. Fig. 2. SEM micrograph of the cross section acquired from the Ti interlayer with a thin graphite substrate. Fig. 3. SEM images and EDS spectra of the cross-section acquired from graphite substrates with Ti interlayer in three divided areas at four deposition times: (a) 1 h, (b) 26 h, (c) 72 h, (d) 152 h; (1) area 1, (2) area 2, (3) area 3. Fig. 4. X-ray diffraction spectra (XRD) of samples in three disparate, divided areas at four deposition times: (a) 1 h, (b) 26 h, (c) 72 h, (d) 152 h; (1) area 1, (2) area 2, (3) area 3. Fig. 5. Evolution process of the Ti interlayer at various deposition times. Fig. 6. Residual stress of diamond films in each divided area for three deposition times: (a) 26 h, (b) 72 h, (c) 152 h. Fig. 7. Macrograph of an as-deposited diamond film with typical cracks.
Fig. 1. Static state image of the arc and arc distributions of the DC arc plasma jet CVD.
Fig. 2. SEM micrograph of the cross section acquired from the Ti interlayer with a thin graphite substrate.
Fig. 3. SEM images and EDS spectra of the cross-section acquired from graphite substrates with Ti interlayer in three divided areas at four deposition times: (a) 1 h, (b) 26 h, (c) 72 h, (d) 152 h; (1) area 1, (2) area 2, (3) area 3.
Fig. 4. X-ray diffraction spectra (XRD) of samples in three disparate, divided areas at four deposition times: (a) 1 h, (b) 26 h, (c) 72 h, (d) 152 h; (1) area 1, (2) area 2, (3) area 3.
Fig. 5. Evolution process of the Ti interlayer at various deposition times.
Fig. 6. Residual stress of diamond films in each divided area for three deposition times: (a) 26 h, (b) 72 h, (c) 152 h.
Fig. 7. Macrograph of an as-deposited diamond film with typical cracks.
S0169-4332(16)30336-1 http://dx.doi.org/doi:10.1016/j.apsusc.2016.02.158 APSUSC 32676
To appear in:
APSUSC
Received date: Accepted date:
2-12-2015 18-2-2016
Please cite this article as: Jianchao Guo, Chengming Li, Jinlong Liu, Junjun Wei, Liangxian Chen, Chenyi Hua, Xiongbo Yan, Structural evolution of Ti destroyable interlayer in large-size diamond film deposition by DC arc plasma jet, Applied Surface Science http://dx.doi.org/10.1016/j.apsusc.2016.02.158 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Structural evolution of Ti destroyable interlayer in large-size diamond film deposition by DC arc plasma jet Jianchao Guo, Chengming Li, Jinlong Liu, Junjun Wei, Liangxian Chen, Chenyi Hua, Xiongbo Yan Institute of Advanced Materials and Technology, University of Science and Technology Beijing, Beijing 100083, P.R. China Graphical abstract
Highlights 1 The evolution processes of a Ti modeled interlayer were evaluated in DC arc plasma. 2 Sandwich structures were observed along with the growth of TiC. 3 Ti interlayer released stress via self-fracture.
Abstract: The addition of titanium (Ti) interlayer was verified to reduce the residual stress of diamond films by self-fracturing and facilitate the harvest of a crack-free free-standing diamond film prepared by direct current (DC) arc plasma jet. In this study, the evolution of the Ti interlayer between large-area diamond film and substrate was studied and modeled in detail.
Corresponding author. E-mail address: [email protected] (C. M. Li)
The evolution of the interlayer was found to be relevant to the distribution of the DC arc plasma, which can be divided into three areas (arc center, arc main, and arc edge). The formation rate of titanium carbide (TiC) in the arc main was faster than in the other two areas and resulted in the preferred generation of crack in the diamond film in the arc main during cooling. Sandwich structures were formed along with the growth of TiC until the complete transformation of the Ti interlayer. The interlayer released stress via self-fracture. Avoiding uneven fragile regions that formed locally in the interlayer and achieving cooperatively released stress are crucial for the preparation of large crack-free diamond films. Keywords: free-standing diamond film; DC arc plasma jet; Ti destroyable interlayer; evolution process
1. Introduction Polycrystalline diamond film has several exceptional properties that make it an outstanding optical material or other functional materials [1–3]. The demand for diamond films has increased in many applications, especially for large-area diamond films [4,5]. Among all deposition methods for large-area polycrystalline diamond film, direct current (DC) arc plasma jet chemical vapor deposition (CVD) is one of the most popular techniques because it can prepare large-area free-standing diamond films with a high deposition rate [6–8]. In the DC arc plasma jet CVD, the high-energy activation process by DC voltage leads to increased plasma temperature for C–H or C–H–O radicals. A higher deposition rate can be reached compared with that through microwave CVD or other methods. Nevertheless, synthesizing a large freestanding diamond film without crack remains a challenge because of the different heat expansion coefficients between prepared diamond films and the substrate. Large residual stress in diamond films is undoubtedly the primary cause of diamond film fractures [9]. The low success ratio of crack-free diamond films restricts the development and application of largearea diamond films. The introduction of an interlayer is an effective method to reduce the residual stress of diamond films [10–12]. Among all conventional interlayer materials, titanium (Ti) is the primary element employed extensively for diamond film preparation. Ti can release the stress of a diamond film via self-fracture, and hence improve the success rate of a crack-free diamond film [9,13,14]. Prior to diamond deposition, the titanium carbide (TiC) layer is formed first, which is important for diamond nucleation and growth [15]. In particular, the conversion degree from Ti interlayer to TiC is affected by the deposition environment. For example, the TiC layer
can be as thick as 150 μm or too thin to be detected under different conditions [16,17]. As for the DC arc jet CVD method, the DC arc plasma distributions above the substrate surface can be divided into three areas, namely, arc center, arc main, and arc edge [18]. These areas have an impact on TiC formation. In the present paper, polycrystalline diamond films (120 mm in diameter) were deposited on the graphite substrate with a Ti interlayer by DC arc plasma jet CVD operating at the gas recycling mode. The correlations among the evolution process of Ti interlayer, distributions of residual stress, and typical fracture mode in large-size diamond films were analyzed and discussed by detecting the variation of microstructures and compositions of Ti interlayer in different divided areas with various deposition times. 2. Experimental Polycrystalline diamond films (120 mm in diameter) were deposited on the Ti/graphite substrate via DC arc plasma jet. The equipment and processes had been discussed previously [9,19]. The Ti interlayer (approximately 8 μm thick) was deposited using multi-arc ion plating equipment. The diamond films were deposited for different deposition times (i.e., 1, 26, 72, and 152 h). The process parameters used in this study were as follows: power, 28–31 KW; substrate temperature, 800–950 ℃; chamber pressure, 2.5–3.8 KPa; CH4 concentration, 1 %. Apart from that in the nucleation stage, the deposition rate of diamond films was about 10 μm/h. The substrate surface was divided into three areas according to arc characteristics. After deposition of the diamond film, the graphite layers of samples were gently ground to a thickness of 0.1 mm to further confirm the correlation between the typical fracture mode of large diamond films and the unsynchronized evolution of the Ti interlayer caused by arc
characteristics. Scanning electron microscopy (SEM) of the QUANTAFEG 250 system was used to examine the morphology of the interlayer cross section. The distributions of C and Ti across the interface were evaluated through backscattered electron mode. Changes in Ti and C concentrations on the cross section of samples were evaluated using the line scan pattern energy dispersive spectrometer (EDS) available in the SEM system. The SMARTLAB X-ray diffraction (XRD) equipment was used to confirm the presence of TiC in the Ti interlayer. Furthermore, to clearly show the effect of the Ti interlayer on stress, the XRD sin2ψ method was used to statistically determine the residual stress in CVD diamond films deposited on graphite substrates with a Ti interlayer. The tests were conducted from the nucleation surface of the diamond because the diamond films became thick after deposition for hundreds of hours, and a discrepancy in growth and nucleation sides of the diamond films was found [20]. 3. Results and discussion 3.1. Arc characteristic of DC arc plasma jet The DC arc plasma generated between cathode and circle anode has a point-to-point discharge pattern. The arc plasma driven by external magnetic fields and rotated gas is in highspeed rotating mode, which can reach several thousand rpm. Given that the arc plasma is affected by electric, magnetic, and flow fields, the reaction is extremely complicated. In this study, the static image of DC arc plasma was obtained using a high-speed camera (Fig. 1). On the basis of the shape characteristics, arc distributions above the surface space of a substrate can be divided into arc center, arc main, and arc edge [18]. Correspondingly, three areas on the substrate surface were labeled as area 1, area 2, and area 3. The arc main is closer to the
substrate compared with the other two regions. The special shape of the arc situated the Ti interlayer on the substrate in a non-homogeneous environment, which caused different evolution processes for Ti interlayer in different areas. 3.2. Evolution process of the Ti interlayer 3.2.1. Changes in sectional morphology characteristics and element concentration distributions Fig. 2 shows the columnar grains in the Ti interlayer before the deposition of the diamond film. The sectional morphology characteristics and element concentration distributions of the Ti interlayer in three divided areas at four deposition times are shown in Fig. 3. The fracture morphology from the top of the diamond, Ti transition layer, and graphite substrate is also displayed. Figs. 3(a1)–(a3) show that the Ti interlayer began to gradually transfer to the isometric grain, and the columnar crystal nearly disappeared after 1 h of diamond film deposition. During the first 26 h, the interlayer thickened further and fine-grain layers appeared on the interfaces of the Ti interlayer. The sandwich structure was formed, and no significant differences among the three different areas were found, as shown in Figs. 3(b1)–(b3). The fine grains met in the center of the Ti layer within 72 h in areas 1 and 2, as shown in Figs. 3(c1) and 3(c2) respectively. The grain size of the Ti interlayer in area 2 was clearly smaller than that in area 1 from 72 h to 152 h, which proved that the extending rate in area 1 was slower than in area 2. The growth rate of the fine-grain layer in area 3 was significantly lower than that in the other two areas with prolonged deposition times. From Figs. 3(d2)–(d3), numerous cracks were found in the fine-grain-enriched regions when
deposition time was over 152 h, especially in area 2. Thus, the release of stress destroyed the interlayer. The cracks could be divided into horizontal and vertical directions according to the propagation orientation of cracks. Horizontal cracks appeared on the center of the interlayer (the intersectional region of small grains from two sides) and the junction regions of Ti layer and two carbon layers (diamond layer and graphite layer). The positions of the vertical cracks were stochastic. Area 2 is taken as an example to determine the relationship between fine grains and element distributions with different deposition times. The element distributions in the cross–section depth of samples were evaluated by line scanning of the EDS spectrum. The results are shown in Figs. 3(a2)–(d2). At the preliminary stage of diamond film deposition for 1 h [Fig. 3(a2)], no obvious diffusion of Ti and C could be found in the interlayer. Fig. 3(b2) shows that when deposition times increased to 26 h, fine-grain layers appeared on both sides of the Ti layer. Correspondingly, a dramatic diffusion of Ti and C proceeded in both interfaces. As deposition time increased to 72 h, the C element clearly distributed in the interlayer, and the carbon content was relatively high in small-grain regions, as shown in Fig. 3(c2). Figs. 3(c2) and 3(d2) show that only weak variations of the concentration distributions were found when the deposition times increased from 72 h to 152 h. Changes in C concentration were synchronized with the extension of fine grains shown in the SEM results. Thus, the fine grains were probably the carbonization products of the Ti interlayer. 3.2.2. Changes in X-ray diffraction spectra Fig. 4 shows the X-ray diffraction spectra of the samples coming from three divided areas of nucleation surface of diamond films deposited for different time periods. Clear diffraction
peaks of Ti, TiC, TiO, diamond, and graphite are shown in three divided areas for 1 h of deposition [Figs. 4(a1)–(a3)]. TiO should be generated during coating or chemical reaction with air before diamond film deposition. The TiC between diamond and Ti interlayer was formed because of the carbon atom in the deposition atmosphere defusing into the Ti layer and between graphite and the Ti interlayer from substrate. With increased deposition times, diffraction peaks of the diamond in every area were significantly enhanced, and TiC diffraction peaks were much more apparent in areas 1 and 2 [Figs. 4(b1)–(b3)]. Ti diffraction peaks still appeared in 26 h and 72 h of deposition, as shown in Figs. 4(b1)–(c3). These peaks corresponded to the sandwich structure in Fig. 3. TiC diffraction peaks were enhanced, and Ti diffraction peaks disappeared when deposition times increased from 72 h to 152 h in Figs. 4(d1)–(d3). This result demonstrated the completion of the carbonization process to some degree. 3.2.3. Model and discussion of the evolution process of Ti interlayer The evolution process of the Ti interlayer is schematically displayed in Fig. 5. Fig. 5(a) shows that the columnar grains in the Ti interlayer are exhibited before diamond film deposition. Fig. 5(b) shows that columnar crystals have a declining trend after nucleation, which may be caused by recrystallization at high temperature. The development of sandwich structures (Fig. 5c) causes TiC to come into contact with both sides of the interlayer. A dividing line can also be seen in Figs. 3(d1)–(d3). The crack at the completion of diamond film deposition may have originated from the interfaces of TiC/TiC in the interlayer, diamond/TiC, and graphite/TiC. The significance of the destroyable interlayer is demonstrated in Figs. 3(d1)– (d3) and Fig. 5(d). In addition, given the elasticity modulus of diamond, TiC, and graphite at 700–1200 GPa, 448–451 GPa, and 4.8–28 GPa, respectively, the stress buffer is formed during
cooling because of the gradient of elasticity modules. Under the increased stress conditions, the interlayer uses self-fracture to release stress during cooling after diamond deposition. 3.3. Residual stress distributions of diamond films The residual stress distributions in diamond films after cooling were inevitably influenced by the uneven evolution of the Ti interlayer. The XRD sin2ψ method was used to determine residual stress in CVD diamond films for different deposition time periods in three divided areas. The (311) plane of CVD diamond was used with a tilt angle (ψ) from 0° to 45°. To guarantee the reliability of detection results, three points in each divided area were detected. The statistical residual stress distributions of diamond films in divided areas for different deposition times are shown in Fig. 6. Given that the diamond grains had not yet formed a complete film, the samples deposited for 1 h were not considered. With the increase of deposition time, the thickness of diamond films after different deposition times were about 250, 710 and 1500 μm, respectively. There was no obvious thickness difference of diamond films in three different areas. The thermal expansion coefficient mismatch among the interlayer and diamond film caused large compressive stress states to be detected in all areas of the diamond film after different deposition times. Fig. 6(a) shows that, during 26 h of deposition, the magnitudes of residual stress in three areas showed no obvious difference and each standard deviation was relatively small. Stress distributions ranging from –1.8 GPa to –2.15 GPa followed a slightly increasing trend from area 1 to area 3. Fig. 6(b) shows that after deposition for 72 h, the average stress of diamond films was slightly increased and stress distribution trends remained unchanged. However, the standard deviations of stress results in area 2 were significantly higher than in
the other two areas. Fig. 6(c) shows that when deposition times reached 152 h, the residual stress in areas 2 and 3 were reduced. The average stress in area 2 was –1.55 GPa, which was the lowest among three divided areas. SEM results showed more cracks in area 2 than in the other two areas, which reduced stress and caused uneven distributions of stress in diamond films, and a large stress gradient between areas 1 and 2. In summary, with less than 72 h of deposition time, only a part of the Ti interlayer transformed into TiC on both sides of Ti/graphite and diamond/Ti. The Ti layer had a low elastic modulus, and thus elastic deformation of the interlayer would gradually increase the compressive stress from center to edge in diamond films after cooling. When the deposition time was long enough, almost all the Ti interlayer transformed into TiC and more stress-fragile regions in area 2 were found. Achieving elastic deformation is difficult because of the relatively high elastic modulus of TiC. Thus, the interlayer in area 2 was first destroyed during the cooling process and stress was drastically released. 3.4. Typical cracks in diamond films Fig. 7 presents the typical cracks in diamond films (120 mm in diameter) deposited for almost 150 h. Main ring cracks were found in area 2, and parts of the cracks extended to area 3. An analysis of the characteristics of typical cracks indicated that the fracture mode is likely related to Ti interlayer features caused by arc area characteristics. Diamond films in area 2 had the largest residual stress drop, and more cracks correspondingly appeared in the interlayer. Thus, stress-releasing behavior occurred in area 2 due to the fragile regions formed in the Ti interlayer. Theoretical simulation showed that the temperature in area 1 was the highest, which decreased with increased distance to area 1 [21]. The special “L”-type rotating arc and the
intake method of carbon gases caused the highest carbon concentration to be achieved in area 2. Relatively high temperature and the highest carbon concentration were observed near the substrate surface in area 2. The transformation rate was the fastest and resulted in more fragile regions appearing in the interlayer of area 2. Thus, area 2 was not conducive to cooperative stress release in different areas during cooling. As described above, the unsynchronized transformation of the Ti interlayer caused uneven distributions of residual stress and an increased fracture probability in diamond films, especially in area 2. Thus, cooperative stress release of interlayer in different areas was crucial for preparing large crack-free diamond films. 4. Conclusions The introduction of the Ti interlayer effectively reduces the residual stress of diamond films in high-power DC arc plasma jet CVD operating at the gas recycling mode. With increased deposition times, the TiC layer formed on both sides of the Ti interlayer and extended inward at different rates. Sandwich structures were formed along with TiC growth until the complete transformation of the Ti interlayer. The unsynchronized evolution of the Ti interlayer caused uneven distributions of residual stress and led to typical ring cracks in large-size diamond films.
Acknowledgments This work was sponsored by the National Natural Science Foundation of China (No.51272024) and the Ph.D. Programs Foundation of the Ministry of Education of China (No.20110006110011). The authors deeply appreciate their financial support.
References [1] S. Coe, R. Sussmann, RS, Optical, thermal and mechanical properties of CVD diamond, Diamond Relat. Mater. 9 (2000) 1726-1729. [2] R. Sussmann, J. Brandon, G. Scarsbrook, C. Sweeney, T. Valentine, A. Whitehead, C. Wort, Properties of bulk polycrystalline CVD diamond, Diamond Relat. Mater. 3 (1994) 303-312. [3] A. Gicquel, K. Hassouni, F. Silva, J. Achard, CVD diamond films: from growth to applications, Curr. Appl. Phys. 1 (2001) 479-496. [4] A. Kromka, O. Babchenko, T. Izak, K. Hruska, B. Rezek, Linear antenna microwave plasma CVD deposition of diamond films over large areas, Vacuum. 86 (2012) 776-779. [5] J. Weng, L. Xiong, J. Wang, S. Dai, W. Man, F. Liu, Investigation of depositing large area uniform diamond films in multi-mode MPCVD chamber, Diamond Relat. Mater. 30 (2012) 15-19. [6] J.V. Busch, J.P. Dismukes, Trends and market perspectives for CVD diamond, Diamond Relat Mater. 3 (1994) 295-302. [7] S. Matsumoto, Y. Manabe, Y. Hibino, Diamond deposition using anX-Y stage in a dc plasma jet chemical vapour deposition, J. Mater. Sci. 27 (1992) 5905-5910. [8] N. Ohtake, M. Yoshikawa, Nucleation effects and characteristics of diamond film grown by arc discharge plasma jet chemical vapor deposition, Thin solid films. 212 (1992) 112121. [9] C. Li, L. Wang, L. Chen, Z. Liu, L. Hei, F. Lu, Free-standing diamond films deposited by DC arc plasma jet on graphite substrates with a destroyable Ti interlayer, Diamond Relat. Mater. 18 (2009) 1348-1352. [10] R. Polini, F.P. Mantini, M. Barletta, R. Valle, F. Casadei, Hot filament chemical vapour deposition and wear resistance of diamond films on WC-Co substrates coated using PVDarc deposition technique, Diamond Relat. Mater. 15 (2006) 1284-1291. [11] X. Peng, T. Clyne, Formation and adhesion of hot filament CVD diamond films on titanium substrates, Thin Solid Films. 293 (1997) 261-269. [12] Y. Li, Y. Tang, Q. Yang, C. Xiao, A. Hirose, Diamond deposition on steel substrates with an Al interlayer, Int. J. Refract. Met. Hard Mater. 27 (2009) 417-420. [13] Y. Zhu, L. Wang, W. Yao, L. Cao, Interface diffusion and reaction between Ti layer and Si3N4/Si substrate, Surf. Interface Anal. 32 (2001) 296-300. [14] E. Buccioni, E. Braca, J. Kenny, M. Terranova, Processing-structure-adhesion relationship in CVD diamond films on titanium substrates, Diamond Relat. Mater. 8 (1999) 17-24. [15] T. Grogler, E. Zeiler, R. Singer, S. Rosiwal, M. Dannenfeldt, Growth stages of chemical vapor deposited diamond on the titanium alloy Ti-6Al-4V, Diamond Relat. Mater. 6 (1997) 1658-1667. [16] Q. Chen, Z. Lin, Diamond growth on thin Ti wafers via chemical vapor deposition. J, Mater. Res. 10 (1995) 2685-2688. [17] S.S. Park, J.Y. Lee, Synthesis of diamond films on titanium substrates by hot-filament chemical vapor deposition, Appl. Phys. 69 (1991) 2618-2622. [18] C. Li, R. Zhu, J. Liu, L. Chen, J. Guo, C. Hua, L. Hei, J. Wei, J. Wang, Z. Feng, Effect of
arc characteristics on the properties of large size diamond wafer prepared by DC arc plasma jet CVD, Diamond Relat. 39 (2013) 47-52. [19] F. Lu, G. Zhong, J.G. Sun, Y. Fu, W. Tang, J. Wang, G. Li, J. Zang, C. Pan, C. Tang, A new type of DC arc plasma torch for low cost large area diamond deposition, Diamond Relat. Mater. 7 (1998) 737-741. [20] R. Zhu, C. Li, L. Chen, J. Liu, J. Wei, L. Hei, H. Guo, Erosion difference of growth and nucleation sides of free-standing diamond films prepared by DC arc plasma jet CVD, Appl. Surf. Sci. 355 (2015) 203-208. [21] T. Huang, W. Tang, F. Lu, J. Gracio, N. Ali, Argon-to-hydrogen ratio in plasma jet diamond chemical vapour deposition, Surf. Coat. Technol. 190 (2005) 48-53.
Figure captions Fig. 1. Static state image of the arc and arc distributions of the DC arc plasma jet CVD. Fig. 2. SEM micrograph of the cross section acquired from the Ti interlayer with a thin graphite substrate. Fig. 3. SEM images and EDS spectra of the cross-section acquired from graphite substrates with Ti interlayer in three divided areas at four deposition times: (a) 1 h, (b) 26 h, (c) 72 h, (d) 152 h; (1) area 1, (2) area 2, (3) area 3. Fig. 4. X-ray diffraction spectra (XRD) of samples in three disparate, divided areas at four deposition times: (a) 1 h, (b) 26 h, (c) 72 h, (d) 152 h; (1) area 1, (2) area 2, (3) area 3. Fig. 5. Evolution process of the Ti interlayer at various deposition times. Fig. 6. Residual stress of diamond films in each divided area for three deposition times: (a) 26 h, (b) 72 h, (c) 152 h. Fig. 7. Macrograph of an as-deposited diamond film with typical cracks.
Fig. 1. Static state image of the arc and arc distributions of the DC arc plasma jet CVD.
Fig. 2. SEM micrograph of the cross section acquired from the Ti interlayer with a thin graphite substrate.
Fig. 3. SEM images and EDS spectra of the cross-section acquired from graphite substrates with Ti interlayer in three divided areas at four deposition times: (a) 1 h, (b) 26 h, (c) 72 h, (d) 152 h; (1) area 1, (2) area 2, (3) area 3.
Fig. 4. X-ray diffraction spectra (XRD) of samples in three disparate, divided areas at four deposition times: (a) 1 h, (b) 26 h, (c) 72 h, (d) 152 h; (1) area 1, (2) area 2, (3) area 3.
Fig. 5. Evolution process of the Ti interlayer at various deposition times.
Fig. 6. Residual stress of diamond films in each divided area for three deposition times: (a) 26 h, (b) 72 h, (c) 152 h.
Fig. 7. Macrograph of an as-deposited diamond film with typical cracks.