Morphology and mechanical behavior of diamond films fabricated by IH-MPCVD

Abstract

Morphology of diamond films has been controlled via intermediate frequency induction heated microwave plasma chemical vapor deposition (IH-MPCVD), which was transformed with various substrate temperatures (T_sub = 923–1123 K) and CH₄/H₂ ratios (η_c = 0.5–2 vol%). The coupling effects of T_sub and η_c on the structure of diamond films have been studied. At η_c = 0.5 vol%, the sp³/sp² ratio of diamond films reached 98% at 1073 K, surface roughness (R_ms) increased from 50 to 85 nm with increasing T_sub, the maximum hardness (H_a) reached 84 GPa at 973 K, and the maximum Young's modulus (E) reached 642 GPa at 1023 K. The residual stress (σ) was calculated as a function of T_sub and η_c. The quality factor (Q), combining microstructure and mechanical behavior, has been creatively defined to evaluate the quality of diamond films.

---

1. Introduction

Chemical vapor deposited (CVD) diamond covers a broad range of synthetic diamond materials with great variety of properties such as high hardness, high elastic modulus, chemical inertness, and low self-friction coefficient.^1–4 CVD diamond can be broadly divided into two grades, i.e., microcrystalline diamond (MCD) and nanocrystalline diamond (NCD), with the crystallites in the micron size range and smaller than 500 nm, respectively.⁵ NCD films are highly desirable for the substrates of MEMS (Micro-Electro-Mechanical Systems) because the smaller grain size insures a smoother surface for the devices above comparing with MCD.^6–8 However, the mechanical properties of NCD, such as hardness and elastic modulus, are lower than those of conventional MCD due to the presence of sp² non-diamond species.⁹ Therefore, the smooth surface, good mechanical property and high sp³/sp² ratio are usually difficult to obtain simultaneously, and it is hard to evaluate the quality of diamond-related materials only according to their individual properties. Here, we define quality factor (Q), combining microstructure and mechanical behavior to evaluate the deposits.

Besides, substrate temperature (T_sub) plays an important role in growth process of diamond. Not only the structure and properties of CVD diamond films, but also the growth rate is mainly influenced by T_sub. For example, the commonly low T_sub may hinder the formation of strong chemical bonds between atoms, while the high temperature may enhance the graphite species.¹⁰ However, it is difficult to control precisely the T_sub in MPCVD system because the T_sub is attributed to several factors, e.g., discharging plasma density and substrate location, which is influenced strongly by the microwave power (p_w). Moreover, T_sub would certainly decrease due to the heat exchange between the chamber and external environment. And it is hard to increase the decreased T_sub by adjusting p_w due to its strong influence on T_sub. In general, T_sub is difficult to be precisely controlled only through the adjustment of p_w, which resulting in the difficulty of researches on diamond structures.

In this study, an IH-MPCVD apparatus was developed to fabricate diamond films. The T_sub was compensated with an intermediate frequency induction heating system by self-adjusting the IH power (p_h) according to the thermocouple that set under the substrate. The effects of CH₄/H₂ ratio and deposition temperature on the morphology, quality factor (Q) and deposition rate of the diamond films has been investigated, and the growth mechanism of these films was also proposed.

2. Experimental

Fig. 1 shows the experimental schematic of the home-made induction heated microwave plasma CVD (IH-MPCVD). The system consisted of a microwave generator network (MPS-15D, NISSIN, Tokyo, Japan, 2.45 GHz, 1.5 kW), a quartz tube chamber, an intermediate frequency induction heating device (S08-6093, Daiichi Kiden, Tokyo, Japan, 6 kHz, 12.5 kW), and a vacuum system. A hollow graphite cylinder heated by the induction heating device was set below the substrate holder as heating unit surrounded by a graphite felt. The substrate holder was supported by an alumina tube within a K-type thermocouple. The temperature signals of thermocouple were fed back to the IH-system.

Fig. 1 Schematic diagram of the IH-MPCVD system.

Single-crystalline silicon 〈100〉 wafers with dimensions of 10.0 × 15.0 × 0.5 mm were used as substrates. To enhance nucleation, the Si substrates were scratched via ultrasonic bath in acetone with diamond powders (10–40 μm) for 1 hour. The scratched substrates were taken ultrasonic cleaning in deionized water, and then dried in a N₂ flow.

The substrates were preheated at 873 K in 30 minutes after quartz chamber was evacuated to 10⁻¹ Pa. The input gas flow rate of H₂ (f_h) was 400 sccm. The flow rate of CH₄ (f_c) was varied from 1 to 10 sccm. The total pressure (P_tot) and the microwave power (p_w) were 4 kPa and 900 W, respectively. The induction heating power (p_h) of the intermediate frequency induction heating device was self-adjusted from 0 to 12.5 kW, according to the substrate temperature assessed by thermocouple. Thus, the substrate temperatures (T_sub) were controlled in the range of 923–1123 K. The depositions for all samples were carried out for a total time of 3 h.

Phase identification of the films was obtained by Raman spectra (LabRAM HR Evolution; Horiba, Paris, France) with the excitation of a diode laser 532 nm in wavelength. Crystalline phases were examined by X-ray diffraction with Cu-Kα radiation (XRD; Ultima III, Rigaku, Tokyo, Japan, at 40 kV and 40 mA). A field-emission scanning electron microscope (SEM; Quanta-250, FEI, Houston, TX, at 20 kV) was used to observe the film thickness and microstructure. The surficial roughness was analyzed by atomic force microscopy (AFM; Multimode 8-HR, Bruker, Santa Barbara, USA). The hardness and young's modulus were analyzed by a MTS Nano-indenter (Agilent Technologies G200, California, USA).

3. Results and discussion

Fig. 2(a)–(e) shows the surficial morphology of diamond films grown at T_sub = 1073 K and various CH₄/H₂ ratio (η_c) ranging from 0.5 to 2.5 vol%. At η_c = 0.5%, the film consisted of regulated quadrate grains with average size of 0.36 μm [Fig. 2(a)]. At η_c = 1.0% and 1.5%, the films exhibited faceted columnar structure with average grain size of 0.34 μm and 0.44 μm, respectively [Fig. 2(b and c)]. At η_c = 2.0%, the grains were coated by nanoparticles [Fig. 2(d)]. By increasing η_c to 2.5%, the films showed dense spherical “cauliflower-like” morphology [Fig. 2(e)]. The cauliflower-like structures was also observed by Hemawan¹¹ via AP-MPCVD. They explained that the formation of spherical and polygonal shapes in the atmospheric pressure MPCVD was due to surface diffusion and condensation from vapor phase.¹² The cross-sectional view [Fig. 2(f–j)] revealed the film thickness reached the maximum at η_c = 1.5%.

Fig. 2 Surficial and cross-sectional SEM image of diamond films deposited at T_sub = 1073 K and various CH₄/H₂ ratio (η_c): (a and f) 0.5, (b and g) 1.0, (c and h) 1.5, (d and i) 2.0 and (e and j) 2.5 vol%.

The minimum roughness (R_ms) of film surface is approximately 23.4 nm at η_c = 2.0–2.5%, as shown in Fig. 3, which is much smaller than that obtained at low η_c = 1.0% (about 80 nm). Fig. 4 shows the XRD patterns of the deposit, the peaks corresponding to diamond (111) at 43.9° (2θ) were only detected for all the specimens, indicating (111) orientation. In terms of carbon chemistry in the plasma discharge, relatively higher η_c results in rich [C_xH_y] and other methyl radicals which are important for diamond nucleation. These species enhance the re-nucleation process of diamond, as well as the sp² hybridized carbon formation.¹³ By reducing the CH₄ fraction in the plasma, the C_xH_y species decreased, especially the species with x > 3 (e.g. [C₄H₂], [C₃H₂]). As a result, the hydrocarbon species with x ≤ 2 incorporates into the lattice by forming a C–C bond to an atom that belongs to a lower-lying terrace, and weaken the nucleation of sp² hybridized carbon phase.⁵ Therefore, well-faceted diamond prefers relatively lower η_c under a constant microwave power in this study.

Fig. 3 Effect of η_c on the surface roughness and growth rate of the diamond films at T_sub = 1073 K.

Fig. 4 XRD patterns of films deposited at T_sub = 1073 K and various η_c.

Fig. 5 shows the cross-sectional and surficial SEM image of diamond films deposited at various T_sub ranging from 923 to 1123 K at η_c = 0.5%. The cross-sectional images revealed that the film thickness varied from 0.5 to 1.9 μm [Fig. 5(f)–(j)]. Namely, in this range, the substrate temperature caused an enhancement in diamond growth rate (R_G).

Fig. 5 Cross-sectional and surficial SEM image of diamond films deposited at T_sub = 923 (a and f), 973 (b and g), 1023 (c and h), 1073 (d and i) and 1123 K (e and j) and η_c = 0.5 vol%. (k–n) is the Wulff configurations for diamond as a function of the facet growth ratio of {100} over {111}.

The magnified details of the samples are given in Fig. 5(a)–(e), pyramidal micrometric diamond crystals are visible at the terminal surface, both triangular and square facets corresponding to {111} and {100} lattice plane were observed. The quadrate planes of {100} facets appear at all the samples [Fig. 5(a–e)], while the triangular {111} facets present at lower temperature [Fig. 5(a–c)]. Area ratio of square and triangular planes obviously grew larger with the increasing of T_sub (923–1073 K), owing to the different growth ratio (R_hkl) of each plane. According to the Wulff configurations,¹⁴ the slowest growing plane would be the one survive in growing diamond particles, and vice versa, the fastest one would disappear completely. Fig. 5(k–n) illustrate the Wulff configurations for diamond deposited at 923, 973, 1023, 1073 K respectively as a function of the facet growth ratio of {100} over {111}, defined as parameter α. The XRD result of this group sample shows that only diamond (111) at 43.9° (2θ) was detected, indicating (111) orientation.

Wild et al.¹⁵ introduced the growth parameter α which relates the growth rates along different crystal axes; 〈111〉 direction would be the fastest growing direction when α is 1, and the growth morphology would be cubic; when α is 3, the fastest growing direction would be 〈100〉 direction, and the growth morphology would be octahedron. For example, in addition to the twinning process, the CVD diamond would show truncated octahedron morphology in various scale when α is 1.5. In this study, for diamond deposited at 923–1023 K [Fig. 5(k)–(m)], α is 1 to 3, and α increase with T_sub. For diamond deposited at 1073 K [Fig. 5(n)], α is close to 3. In conclusion, the parameter α shows a negative relation with the rising T_sub, indicating that the R₁₁₁ increases faster than the R₁₀₀ with the T_sub [Fig. 5(a)–(d)].

Others have reported the order of calculated energy barriers of different diamond surface planes (i.e. {110} < {111} < {100}).^16–19 Namely, {100} surface planes are the most stable facets of CVD diamond grains among H-rich plasma, and the hydrogen abstraction process on {111} facets run faster than those on {100} facets. The whole growth process of CVD diamond would speed up with increasing T_sub, as a result, the growth rate of 〈111〉 direction would be much faster than 〈100〉 direction.

Fig. 6 shows the surficial roughness increased with increasing T_sub except one grown at 1123 K and accompanied by an increasing of film thickness and grain size. These findings are understandable because the diamond film presents a columnar crystal cross-section in Fig. 5, which indicates that the evolution of grain size follows the van der Drift model¹⁴ with columnar structure. The diagram in Fig. 6 depicts the evolution of film thickness and surficial roughness in this temperature series.^6,20,21 The surficial roughness and film thickness showed a positive correlation, except one grown at 1123 K. The AFM topographies suggest that the spherical carbon species may present at higher T_sub, and these amorphous little particles prefer to generate at the ravines between the micro-crystalline grains, which slightly decrease the roughness.

Fig. 6 Evolution of the surface roughness and growth rate of the diamond films synthesized at various T_sub, η_c = 0.5 vol%.

Fig. 7 shows the dependence of the R_G on the T_sub. The R_G of diamond films, with a good quality, is at an intermediate level. The activation energy of 53 kJ mol⁻¹ calculated by Arrhenius plot of the R_G versus a reciprocal T_sub. The result indicates that diamond growth process is limited by temperature controlled chemical process, which under a constant microwave power. Stiegler et al.²⁵ investigated the growth kinetics of diamond films deposited at low substrate temperatures (673–873 K), and they supposed that increasing R_G is most probably caused by the removal of surface bonded hydrogen atoms from the carbon sites. Thus, we induced that with a higher temperature, the removal of surface bonded hydrogen atoms would get easier, and then caused a higher R_G.

Fig. 7 Effect of T_sub on R_G, η_c = 0.5 vol%.

Fig. 8 indicates the Raman spectra for the films deposited at different T_sub. In all samples, a sharp peak has been observed at 1332 cm⁻¹, which corresponds to the zone center of T_2g symmetry of diamond. The other obvious characteristic peak occurs at 1575 cm⁻¹ presenting the zone center E_2g mode of crystalline graphite, and it is usually designated as the “G” band for graphite. Besides, in micro-crystalline graphite, an additional sharp peak should appear at a wave number of 1350 cm⁻¹ in the Raman spectra, which represents a zone-edge A_1g mode, named “D” band.^27,28 However, the peak of “D” band, which close to the sharp diamond T_2g peak, was absent from all the Raman spectra in Fig. 8. Namely, there is little crystalline graphite in these diamond films. In addition to the diamond and graphite, for all the other kinds of amorphous and nanocrystalline carbon species, the Raman spectra typically shows a shoulder and broad “G” band centered around 1550 cm⁻¹ and a “D” band centered at 1375 cm⁻¹.²⁹ The intensity of sharp diamond peak is much higher than the other peaks, which indicates the higher proportion of sp³ carbon atoms in these samples. With increasing T_sub, the related intensity of broad G peak increases, and the “G” band slightly moves up to 1580 cm⁻¹. The slight shift in the broad “G” band frequency is consistent with an increase of amorphous component which produced from the carbon soot formation.³⁰ The downshift and broadening of the diamond peak in nanocrystalline diamond also have previously been observed and explained by quantum confinement effect.³¹ The phase purity is evaluated by the sp³/sp² ratio (r) from the Raman spectra and expressed as eqn (1),³²

where, I_d is the Raman diamond peak area centered at 1332 cm⁻¹ and is the sum of Raman sp² phase peak areas. The factor 75 takes into account the more effective Raman scattering of sp² structures.^33,34 The sp³/sp² ratio of samples shown in Fig. 8 are various from 93 to 98%. However, it should be noticed that two weak peaks occur at 1140 cm⁻¹ and 1450 cm⁻¹, respectively. Others have reported the two peaks not only appear together, yet they also show quite strong and similar dispersions, while the diamond mode does not. Moreover, Kuzmany et al.³⁵ have given a clear proof that these peaks originate from trans-polyacetylene in the grain boundaries by the H–D isotopic substitution result. In Raman features observed here, these two peaks usually occur in the Raman spectra of samples deposited at low temperatures. This is due to the slow growth rate at low T_sub, and the re-nucleating process leads to a high nucleation density with a large amount of boundaries among little grains, where polyacetylene existed. In order to ensure the quality of diamond films, the T_sub during diamond growth process should not be higher than 1073 K, neither be lower than 973 K in this study.

Fig. 8 Raman spectra of the diamond films synthesized at various T_sub, η_c = 0.5 vol%.

As shown in Fig. 9, the surficial morphologies of diamond films can be defined as three categories: the well-faceted crystalline, the spherical structures composited with nanoparticles, and the mixture of both morphologies. At a relative high T_sub (≥1073 K), {100} planes occurred on the film surface. For high CH₄/H₂ ratio (η_c ≥ 2.0%) and low T_sub (≤973 K), more spherical nanoparticles were formed instead of crystalline structure.

Fig. 9 Coupling effects of η_c and T_sub on structure of diamond films.

In order to follow the mechanical properties, a nano-indentation technique was performed. Fig. 10 demonstrates the relationships between average hardness (H_a), Young's modulus (E) and T_sub. The H_a and E were calculated from the correlation between load/unload and displacement (i.e., displacement–load curves not shown here). The theoretical value of Poisson's coefficient (ν) used for calculating the Young's modulus was 0.20, according to the Poisson's ratio of CVD polycrystalline diamond.³⁶ The results reveal that the H_a and E of films increased with T_sub when it ranged in 923–1073 K; the maximum H_a and E reached 84 GPa at 973 K and 642 GPa at 1023 K, respectively. The value is twice as much as regular NCD films by MPCVD,³⁷ much hard than those deposited in argon plasma.³⁸ Exceptionally, both H_a and E decreased marginally when T_sub beyond 1123 K. This anomaly may result from non-diamond phases generated among diamond crystals, confirmed by SEM, AFM and Raman spectra.

Fig. 10 Average hardness (H_a) and Young's modulus (E) of samples fabricated with various T_sub, η_c = 0.5 vol%.

For most applications of diamond films, such as wear-resistant coatings, heatsinks, MEMS and surface acoustic wave devices, a high sp³ component, a high hardness and a low surface roughness are crucial and favorable, but are hard to possess simultaneously. For example, diamond film with an extreme smooth surface always shows a poor mechanical property and low sp³/sp² ratio. To evaluate the quality of different diamond films by microstructure and mechanical behaviors, we proposed a parameter, quality factor (Q), as a formula model:

where, r refers to sp³/sp² ratio, H refers to hardness, R_ms refers to surface roughness, a, b, c are exponential parameters, Q refers to the quality factor of diamond films, the higher the value of Q, the better the quality of diamond films. This formula model is reasonable from a qualitative point of view, because a high r and H, and a low R_ms result in a high value of Q, which means a high quality of diamond film, accordingly, a high sp³ component and hardness, and a low surface roughness are favourable for most applications of diamond films. Here, values of a, b, c were provided as 4, 4, 0.5 respectively, based on the researches in this study after fitting.

Fig. 11 shows Q of DLC,³⁹ NCD⁴⁰ and diamond⁴¹ materials fabricated at different conditions (p_w × η_c × P_tot), in which DLC, NCD and diamond area are classified by Q in the order of 1 × 10⁴, 1 × 10⁵ and 1 × 10⁶, respectively. Through comparing the values of Q, we could judge the quality of the diamond related materials more visually. Most specimens fabricated in this study are in the diamond area and NCD area, which demonstrates that with a relatively good quality, the diamond films in the present study would have a good performance in wear-resistant coatings, heatsinks, MEMS and surface acoustic wave devices.

Fig. 11 Quality factor (Q) of DLC, NCD and Diamond materials fabricated at different conditions (p_w × η_c × P_tot).

The residual stress (σ), as a function of various T_sub at η_c = 0.5% and various η_c at 1073 K was calculated by eqn (3).⁴²

σ = −0.567(ν − ν₀) (GPa) (3)

where ν₀ = 1332 cm⁻¹, ν is the observed Raman shift corresponding to diamond peak, σ < 0 and σ > 0 correspond to the compressive and tensile stress, respectively. The residual stress generally includes intrinsic stress and thermal stress. The thermal stress is due to different thermal expansion coefficients of the film and substrate. Peng et al.⁴³ demonstrated that the intrinsic compressive stress of diamond films was mainly caused by impurities such as sp² bonded carbon and hydrogen, and tensile stress was caused by defects such as excess vacancies, which mainly formed from sp² carbon after etching by atomic hydrogen, and grain boundaries.⁴⁴ Besides, the thicker of diamond film, the higher of the compressive stress.⁴⁵ As shown in Fig. 12, σ increased from −1.54 to −0.79 GPa as T_sub increased from 923 K to 1073 K, which was ascribed to decreasing sp² bonded carbon content with T_sub. However, a decrease of σ was observed at 1123 K due to the presence of spherical carbon species, and increase of film thickness as evidenced by SEM photograph in Fig. 5(e)–(j). With η_c increasing from 0.5 to 2.5%, σ increased from −0.79 to 0.60 GPa. Higher sp² carbon content was formed at higher η_c, most of which was etched away by atomic hydrogen and then became vacancies. The increasing excess vacancies produced tension stress and caused an increase in residual stress. Furthermore, nanocrystalline diamond preferred to grow at relatively higher η_c, as evidenced by SEM photograph in Fig. 2(d and e), resulting in an increase in density of grain boundaries, thus, σ increased.

Fig. 12 Residual stress variation with T_sub (η_c = 0.5%) and η_c (T_sub = 1073 K).

4. Conclusions

To improve the thermal control process of MPCVD technique, we have explored an intermediate-frequency induction heated MPCVD (IH-MPCVD) system. Coupling effects of T_sub (923–1123 K) and η_c (0.5–2 vol%) on structure of diamond films shows that the surficial morphologies of diamond films can be defined as three categories: the well-faceted crystalline, the spherical structures composited with nanoparticles, and the mixture of both morphologies. Higher η_c (2.5 vol%) enhances the re-nucleation process which leads to a smooth surface, yet lower η_c (0.5 vol%) results in better crystalline quality of diamond films. At η_c = 0.5 vol%, the morphology transformed from pyramid to quadrate column with T_sub increasing from 923 to 1073 K, besides, the maximum hardness (H_a) and Young's modulus (E) was 84 GPa and 642 GPa, at T_sub = 1023, 1073 K, respectively. The residual stress increased from −1.54 to −0.79 GPa with T_sub, at η_c = 0.5 vol%, and then decreased, while at 1073 K, it increased with η_c and reached to 0.60 GPa at η_c = 2.5 vol%.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 51372188, and 51521001) and by the 111 Project (B13035). This research was also supported by the International Science & Technology Cooperation Program of China (2014DFA53090) and the Natural Science Foundation of Hubei Province, China (2016CFA006), and the National Key Research and Development Program of China (2017YFB0310400), and the Fundamental Research Funds for the Central Universities (WUT: 2017II43GX, 2017III032, 2017YB004, 2018III016), and Science Challenge Project (No. TZ2016001).

References

1. F. A. Almeida, E. Salgueiredo, F. J. Oliveira, R. F. Silva, D. L. Baptista, S. B. Peripolli and C. A. Achete, ACS Appl. Mater. Interfaces, 2013, 5, 11725.
2. J. J. Gracio, Q. H. Fan and J. C. Madaleno, J. Phys. D: Appl. Phys., 2010, 43, 374017.
3. R. S. Balmer, J. R. Brandon, S. L. Clewes, H. K. Dhillon, J. M. Dodson, I. Friel, P. N. Inglis, T. D. Madgwick, M. L. Markham, T. P. Mollart, N. Perkins, G. a. Scarsbrook, D. J. Twitchen, a. J. Whitehead, J. J. Wilman and S. M. Woollard, J. Phys.: Condens. Matter, 2009, 21, 364221.
4. A. Gicquel, K. Hassouni, F. Silva and J. Achard, Curr. Appl. Phys., 2001, 1, 479.
5. M. Eckert, E. Neyts and A. Bogaerts, Cryst. Growth Des., 2010, 10, 4123.
6. S. H. Seo, W. C. Shin and J. S. Park, Thin Solid Films, 2002, 416, 190.
7. W. C. Shih, M. J. Wang and I. Nan Lin, Diamond Relat. Mater., 2008, 17, 390.
8. Y. Q. Fu, L. Garcia-Gancedo, H. F. Pang, S. Porro, Y. W. Gu, J. K. Luo, X. T. Zu, F. Placido, J. I. B. Wilson, A. J. Flewitt and W. I. Milne, Biomicrofluidics, 2012, 6, 024105.
9. S. A. Catledge, J. Borham, Y. K. Vohra, W. R. Lacefield, J. E. Lemons, S. A. Catledge, J. Borham and Y. K. Vohra, J. Appl. Phys., 2002, 91, 5347.
10. O. A. Williams, M. Daenen, J. D'Haen, K. Haenen, J. Maes, V. V Moshchalkov, M. Nesládek and D. M. Gruen, Diamond Relat. Mater., 2006, 15, 654.
11. K. W. Hemawan, H. Gou, R. J. Hemley, K. W. Hemawan, H. Gou and R. J. Hemley, Appl. Phys. Lett., 2015, 107, 181901.
12. J. Kim, H. Sakakita, H. Ohsaki and M. Katsurai, Jpn. J. Appl. Phys., 2015, 54, 01AA02.
13. H. J. Lee, K. S. Lee, J. M. Cho, T. S. Lee, I. Kim, D. S. Jeong and W. S. Lee, ACS Appl. Mater. Interfaces, 2013, 5, 11631.
14. A. Van der Drift, Philips Res. Rep., 1967, 22, 267.
15. C. Wild, R. Kohl, N. Herres, W. Müller-Sebert and P. Koidl, Diamond Relat. Mater., 1994, 3, 373.
16. J. Bühler and Y. Prior, J. Cryst. Growth, 2000, 209, 779.
17. Y. Zou and K. Larsson, J. Phys. Chem. C, 2016, 120, 10658.
18. K. E. Spear and J. P. Dismukes, in Synthetic Diamond, Wiley, New York, 1994, p. 663.
19. B. T. Van Regemorter and K. Larsson, Chem. Vap. Deposition, 2008, 224.
20. R. Bogdanowicz, M. Sobaszek, J. Ryl, M. Gnyba, M. Ficek, Ł. Gołuński, W. J. Bock, M. Śmietana and K. Darowicki, Diamond Relat. Mater., 2015, 55, 52.
21. F. Bénédic, M. B. Assouar, F. Mohasseb, O. Elmazria, P. Alnot and A. Gicquel, Diamond Relat. Mater., 2004, 13, 347.
22. S. Schwarz, S. M. Rosiwal, M. Frank, D. Breidt and R. F. Singer, Diamond Relat. Mater., 2002, 11, 589.
23. M. S. You, F. C. N. Hong, Y. R. Jeng and S. M. Huang, Diamond Relat. Mater., 2009, 18, 155.
24. M. J. Ulczynski, B. Wright and D. K. Reinhard, Diamond Relat. Mater., 1998, 7, 1639.
25. J. Stiegler, T. Lang, Y. Von Kaenel, J. Michler, E. Blank, J. Stiegler, T. Lang, Y. Von Kaenel, J. Michler and E. Blank, Appl. Phys. Lett., 2014, 173, 88.
26. A. Taylor, L. Fekete, P. Hubík, A. Jäger, P. Janíček, V. Mortet, J. Mistrík and J. Vacík, Diamond Relat. Mater., 2014, 47, 27.
27. P. K. Chu and L. Li, Mater. Chem. Phys., 2006, 96, 253.
28. T. S. Yang, J. Y. Lai, C. L. Cheng and M. S. Wong, Diamond Relat. Mater., 2001, 10, 2161.
29. S. Ray, A. R. Middya and A. K. Barua, Jpn. J. Appl. Phys., 1993, 32, L1559.
30. M. A. Pimenta, G. Dresselhaus, M. S. Dresselhaus, L. G. Cancado, A. Jorio and R. Saito, Phys. Chem. Chem. Phys., 2007, 9, 1276.
31. S. Osswald, V. N. Mochalin, M. Havel, G. Yushin and Y. Gogotsi, Phys. Rev. B, 2009, 80, 75419.
32. F. Silva, A. Gicquel, A. Tardieu, P. Cledat and T. Chauveau, Diamond Relat. Mater., 1996, 5, 338.
33. J. R. Moro, J. Mater. Sci., 2007, 42, 7331.
34. R. E. Shroder, R. J. Nemanich and J. T. Glass, Phys. Rev. B, 1990, 41, 3738.
35. H. Kuzmany, R. Pfeiffer, N. Salk and B. Günther, Carbon, 2004, 42, 911.
36. C. A. Klein and G. F. Cardinale, Diamond Relat. Mater., 1993, 2, 918.
37. W. Kulisch, C. Popov, V. Vorlicek, P. N. Gibson and G. Favaro, Thin Solid Films, 2006, 515, 1005.
38. Y. Wubao, K. Xiang, Y. Size and D. Xiaofeng, Vacuum, 2003, 68, 49.
39. J. S. Hsu, S. S. Tzeng and Y. J. Wu, Vacuum, 2008, 83, 622.
40. T. Sharda, T. Soga, T. Jimbo and M. Umeno, J. Nanosci. Nanotechnol., 2001, 1, 287.
41. M. I. De Barros, L. Vandenbulcke, L. Chinsky, D. Rats and J. Von Stebut, Diamond Relat. Mater., 2001, 10, 337.
42. J. R. Moro, J. Mater. Sci., 2007, 42, 7331.
43. X. L. Peng, Y. C. Tsui and T. W. Clyne, Diamond Relat. Mater., 1997, 6, 1612.
44. H. Windischmann and K. J. Gray, Diamond Relat. Mater., 1995, 4, 837.
45. M. J. Ulczynski, B. Wright and D. K. Reinhard, Diamond Relat. Mater., 1998, 7, 1639.

---