Abstract
Addition of He to a high CH4 content (10.7 vol%) H2/CH4/N2 feedgas mixture for microwave plasma chemical vapor deposition produced hard (56–72 GPa), ultra-smooth nanostructured diamond films on Ti-6Al-4V alloy substrates. Upon increase in He content up to 71 vol%, root mean squared (RMS) surface roughness of the film decreased to 9–10 nm and average diamond grain size to 5–6 nm. Our studies show that increased nanocrystallinity with He addition in plasma is related to plasma dilution, enhanced fragmentation of carbon containing species, and enhanced formation of CN radical.
Diamond films on metal implants provide high hardness, low friction, and wear-resistant coatings, which also are very stable under severe physiological conditions. Smooth nanostructured diamond films with excellent adhesion and hardness up to 80 GPa have been grown on Ti-alloy substrates (Ti-6Al-4V) by microwave plasma chemical vapor deposition (MPCVD) using H2/CH4/N2 feedgas mixtures.1–3 The use of unconventionally high 10–15 vol% CH4 concentrations with addition of 0.1–2 vol% N2 produced diamond films with RMS surface roughness of 15–20 nm and diamond grain sizes of 13–15 nm.
Nanostructured diamond is a composite material which consists of small sp3 nano-crystals of diamond embedded into an amorphous sp2 and sp3 carbon matrix.1–4 Amorphous matrix plays a crucial role in the mechanical, electrical and other properties of nanostructured diamond films. The size of diamond nanocrystals and relative sp2/sp3 content is controlled by plasma chemistry, particularly by the addition of noble gases to the feedgas mixture. Transition from micro- to nanocrystalline diamond film have been observed when very high contents of Ar (≥80 vol%) were added to H2/CH4 plasma.4,5 Significant changes in film morphology were correlated with simultaneous 10–20 fold increase in the optical emission intensity of C2 dimer. The effect of Ar induced nanocrystallinity was explained by the change from CH3 radical diamond growth mechanism at low Ar content to C2 mechanism at high Ar content.4 However, with the exception of Ar, whose influence was studied in detail, addition of other noble gases to feedgas mixture for CVD diamond deposition has been studied by a much lesser extent. Addition of He, Ne, Ar, Kr, and Xe to H2/CH4 plasma produced either polycrystalline diamond films (He, Ar) or graphite films (Ne, Kr, Xe).6 For He/H2/ethanol plasma, the growth of good crystalline diamond film and 3 times increase of growth rate was reported up to 40 vol% He, and poor quality, porous non-diamond films were formed at 70 vol% He.7 In this letter, we report the effect that He addition to H2/CH4/N2 feedgas mixtures has on MPCVD growth of high quality ultra-smooth nanostructured diamond films on Ti-6Al-4V The addition of He significantly reduced film roughness and grain size of diamond nanocrystals without deterioration of film hardness, or adhesive properties.
Nanostructured diamond films of 1.5 μm thickness were deposited by MPCVD on 7 mm diameter Ti-6Al-4V disks, which were initially polished to 4–5 nm RMS surface roughness and treated by ultrasonic agitation in a 1 μm diamond powder/water solution. Total flow rate of He and H2 gases was fixed at 300 sccm and their ratio changed, providing variation from 0 to 71 vol% He. Flow rates of CH4 and N2 were kept constant at 36 and 3.6 sccm (10.7 vol% and 1.1 vol%), respectively. Chamber pressure was 65 Torr and the substrate temperature, as measured by a two-color IR pyrometer was kept in the range 700–740 °C by adjusting microwave power in the range 0.8–0.95 kW. The concentration of plasma species was monitored by optical emission spectroscopy (OES). Glancing angle X-ray diffraction (XRD) with 3° incident beam was used to determine the crystalline structure of the films. The growth rate of diamond film was determined from in situ optical interferometry.
Figures 1a and 1b show AFM images and XRD 2θ-angular dependencies of two films grown with 0 and 71 vol% He. AFM images demonstrate that the film grown with 71 vol% He consists of small 20–30 nm nanoparticles and the film grown without He consists of larger 30–100 nm nanoparticles. In addition, XRD peak of (111) cubic diamond is much broader for the film grown with 71 vol% He. Note, that XRD analysis of all deposited films detected broad (111), (220), and (311) peaks of cubic diamond, and no other carbon related peaks. Significant broadening of diamond peaks upon He addition is related to much smaller average grain size of diamond nanocrystals. The average grain size of diamond was estimated from the full width at half maximum (FWHM) of the (111) diamond peak using Scherrer equation (after correction for instrumental broadening) and presented in Figure 2, together with the surface roughness of the films calculated from 2 ×2 μm AFM images. An increase in He content results in a near linear decrease of diamond grain size from 11–13 nm to 5–6 nm. The surface roughness remains constant at 15–18 nm, or may be slightly increases, up to 30 vol% He, and decreases to 9–10 nm at 71 vol% He. The difference between particle sizes on the AFM image and calculated diamond grain sizes indicates that diamond nanocrystals are agglomerated into larger particles. Addition of He reduces the size of diamond nanocrystals and, probably, the degree of their agglomeration. Micro-Raman spectra for the films at all He contents showed broad peaks at 1137, 1340, 1485, and 1550 cm−1 which are typically observed for nanostructured diamond films containing sp2 and sp3 bonded carbon.1,2 A broad 1340 cm−1 peak is a characteristic of nano-crystalline diamond1–3 and a broad 1550 cm−1 peak is observed in high hardness tetrahedral amorphous carbon films.8 Addition of He made 1340 cm−1 diamond Raman peak broader and did not result in any significant changes in other Raman peaks.
Fig. 1.

1 × 1 μm AFM images and XRD 2θ-angular dependencies of two diamond films grown in He/H2/CH4/N2 plasma with (a) 0 and (b) 71 vol% He.
Fig. 2.

Diamond films grown in He/H2/CH4/N2 plasma at different He contents: (a) FWHM of the (111) diamond XRD peak and calculated average diamond grain size. (b) RMS surface roughness (RMSR) of the films calculated from 2 ×2 μm AFM images.
The hardness and Young’s modulus of the films were measured using a Nanoindenter XP system (MTS Systems, Oak Ridge TN), which was calibrated by using a silica standard. The maximum indentation depth was 150 nm. Nanoindentation showed that the hardness and Young’s modulus of the films do not decrease up to 71 vol% He, and are in the range of 58–72 GPa and 380–480 GPa, respectively.
In order to clarify the mechanism of diamond growth we used OES to monitor changes in plasma chemistry upon He addition. Figures 3a and 3b show the growth rate of diamond film and the ratios of OES intensities of plasma species (CN/Hα, C2/Hα, and Hβ/Hα) as a function of added He. Figure 3a shows a steady drop of the growth rate from 0.63 μm/hr to 0.45 μm/hr up to 71 vol% He. Figure 3b shows that upon He addition the ratio Hβ/Hα remains practically constant, indicating only minor changes in plasma temperature. The ratio C2/Hα increases about 2 times, and the ratio CN/Hα increases about 3 times.
Fig. 3.

(a) Growth rate of diamond films versus He content in He/H2/CH4/N2 plasma. (b) Normalized optical emission intensities of Balmer Hβ (1,486.14 nm), C2 (2,516.5 nm), and CN (3,386 nm) lines versus He content in He/H2/CH4/N2 plasma. Lines were normalized to Balmer Hα line (656.3 nm) intensity.
The effect of He addition on reducing diamond grain size suggests that the rate of secondary nucleation/renucleation increases in He/H2/CH4/N2 plasma, precluding the growth of large diamond nanocrystals. One trivial effect of He addition is simply the increase in CH4/H2 ratio in He/H2/CH4/N2 plasma, which should reduce the effect of hydrogen on suppressing secondary nucleation by regasifying nondiamond carbon. However, our data indicate that the effect of He addition is not a simple effect of plasma dilution, but it is based on a more complex mechanism. Diamond films grown in H2/CH4/N2 plasma without He at corresponding high CH4/H2 ratio of 0.6 have poor quality with high content of graphitic phase. Another explanation of the He effect may be related to the known strong influence of the CN radical on the degree of diamond nanocrystallinity.3 The observed decrease in film roughness above 30 vol% He correlates well with the simultaneous increase in CN/Hα ratio. Nevertheless, CN mechanism alone may not account for observed increase in nanocrystallinity. It has been shown, that a critical N2 content in H2/CH4/N2 plasma, above which CN radical influence on nanocrystallinity is diminished, is lower for higher CH4/H2 ratios.3
Our OES data indicate that the addition of He to H2/CH4/N2 plasma is different from the addition of Ar. Thus the C2/Hα ratio increased 10 times at 70 vol% Ar and only 2 times at similar He content. Even more pronounced is the observed small increase of C2/Hα ratio at very high He contents of 80–98 vol%, compared to its 10–20 times increase at corresponding high Ar contents.5 Thus, we can conclude that the effect of He addition on reducing the diamond grain size cannot be accounted for by the switching from CH3 (or C2H2) growth mechanism to C2 mechanism, which was responsible for formation of nanocrystalline diamond at very high 80–99 vol% Ar.4
It was reported that H atoms in He/H2 plasma may have an extraordinary temperature up to 180–210 eV, which is reflected in Doppler broadening of Hα emission line-width up to 0.6–0.7 nm.11 Consequently, the unique influence of super hot H atoms on the diamond deposition was suggested.6 However, this effect can be excluded for our experimental conditions. Careful investigation of Hα line-width in our deposition chamber did not reveal any Hα line broadening above the instrumental line-width of 0.057 nm under all inspected conditions (pressure of 5–100 Torr, microwave power of 0.6–2 kW, 0–100 vol% He). Probably unusual Hα line broadening is a feature of low pressure (2.5 Torr), low microwave power (40 W) plasma conditions described in the Ref. [10].
The unusual properties of He plasmas have been discussed for deposition of DLC films.11,12 Ionization potential of He is 24.5 eV, which is much higher than that of Ar (15.76 eV). In addition, the excitation energy of long-lived (life-time without quenching is 6 × 105 s) excited state (23S) of He atoms is 19.8 eV, compared to 11.55 eV for much shorter-lived (life-time is 1.3 s) excited (43P2) Ar atom.13 Thus, long-lived energetic excited He atoms can lead to additional ionization and fragmentation of CH4 gas via Penning mechanism.11 From our OES data it is evident that in He/H2/CH4 plasma the fragmentation of C2 dimer is significantly enhanced compared to Ar/H2/CH4 plasma. Enhanced fragmentation of C2 and other carbon containing species in He plasma may suppress the growth of large diamond nanocrystals.
In summary, we have studied the effect of He addition to a high CH4 content (10.7 vol%) H2/CH4/N2 microwave plasma using different He/H2 feedgas ratios. An increase in He content up to 71 vol% resulted in a decrease of diamond grain size from 15 nm to 9 nm. At the same time, surface roughness of the film decreased from 15–18 nm to 9–10 nm. Raman spectra, which were typical for nano-structured diamond films, showed no significant changes upon He addition, with exception of 1340 cm−1 diamond peak broadening. Nanoindentation demonstrated that the hardness and Young’s modulus of the films do not decrease up to 71 vol% He, and are in the range of 58–72 GPa and 380–480 GPa, respectively. From optical emission data we found that the fragmentation of C2 dimer in He-containing plasma is significantly enhanced compared to Ar/H2/CH4 plasma. Thus, the diamond growth by C2 mechanism, which was responsible for nanocrystallinity in 80–99 vol% Ar plasma4,5 is suppressed by He addition. The effect of He addition in reducing diamond grain size and film surface roughness is attributed to plasma dilution, enhanced fragmentation of carbon containing species, and enhanced formation of CN radical.
Acknowledgments
We acknowledge support from the National Institute of Dental and Craniofacial Research under Grant No. R01DE013952. A. M. acknowledges support from the National Science Foundation-Research Experience for Undergraduates (REU) site under Grant No. DMR-0243640.
References and Notes
- 1.Catledge SA, Borham J, Vohra YK, Lacefield WR, Lemons JE. J Appl Phys. 2002;91:5347. [Google Scholar]
- 2.Catledge SA, Vohra YK. J Appl Phys. 1998;84:6469. [Google Scholar]
- 3.Corvin RB, Harrison JG, Catledge SA, Vohra YK. Appl Phys Lett. 2002;84:2550. [Google Scholar]
- 4.Gruen DM. Annu Rev Mater Sci. 1999;29:211. [Google Scholar]
- 5.Zhou D, Gruen DM, Qin LC, McCauley TG, Krauss AR. J Appl Phys. 1998;84:1981. [Google Scholar]
- 6.Mills RL, Sankar J, Voigt A, He J, Ray PC, Dhandapani B. Thin Solid Films. 2005;478:77. [Google Scholar]
- 7.Baranauskas V, Ceragiogli HJ, Peterlevitz AC, Tosin MC, Durant SF. Thin Solid Films. 2000;377:182. [Google Scholar]
- 8.Friedmann TA, Sullivan JP, Knapp JA, Tallant DR, Follstaedt DM, Medlin DL, Mirkarimi PB. Appl Phys Lett. 1997;71:3820. [Google Scholar]
- 9.Afzal A, Rego CA, Ahmed W, Cherry RI. Diam Rel Mater. 1998;7:1033. [Google Scholar]
- 10.Mills RL, Ray PC, Dhandapani B, Mayo RM, He J. J Appl Phys. 2002;92:7008. [Google Scholar]
- 11.Mutsukura N, Miyatany K. Diam Rel Mater. 1995;4:342. [Google Scholar]
- 12.Fedosenko G, Schwabedissen A, Engemann J, Braca E, Valentini L, Kenny JM. Diam Rel Mater. 2002;11:1047. [Google Scholar]
- 13.Chang J, Ishikawa, Kinda H. Atomic and Molecular Ionization Processes in Gases. Tokyo Electric University Press; Tokyo: 1999. [Google Scholar]
