Elsevier

Surface and Coatings Technology

Volume 211, 25 October 2012, Pages 152-157
Surface and Coatings Technology

Toward smooth MWPECVD diamond films: Exploring the limits of the hydrogen percentage in Ar/H2/CH4 gas mixture

https://doi.org/10.1016/j.surfcoat.2011.09.065Get rights and content

Abstract

In Ar-rich Ar–H2–CH4 gas mixture the presence of H2 is found to be beneficial to the plasma stability. On the other hand, too high H2 percentages lead to materials showing a high surface roughness. In the present work, diamond films were grown on p-type Si (100) substrates screening different quantities of H2. The plasma phase and plasma–substrate interface were investigated by in-situ optical emission spectroscopy and pyrometric interferometry to determine the behavior of emitting species and the deposition rates, respectively. The obtained films were characterized by Raman micro-spectroscopy, AFM and SEM techniques. For H2 percentages between 6.3 and 10%, the structure and morphology are characteristic of nanocrystalline films, affording low roughness values when a buffer layer was grown between the diamond coating and the treated silicon surface.

Highlights

► We explore the limits of the hydrogen percentage in Ar–H2–CH4 gas mixtures. ► We develop a new kind of buffer layer between the diamond films and treated silicon substrate. ► We obtain a low roughness of the coatings without modifying the Astex-type reactor.

Introduction

The last two decades have witnessed an increasing interest in diamond film growth. This is essentially due to two reasons: firstly, the unique physical properties of diamond, such as extreme mechanical hardness, highest known thermal conductivity, broad optical transparency from the deep UV to the far IR radiations and chemical inertness, just to cite a few [1], [2]; secondly, the spectacular technical advancement achieved by chemical vapor deposition techniques [3]. Among them, the microwave plasma enhanced chemical vapor deposition (MWPECVD) has proved to be an efficient and convenient strategy for the coating of various substrates. Depending on the experimental conditions, the grain size can range from nanometers, < 10 nm in ultrananocrystalline diamond (UNCD) [4], [5], < 100 nm in nanocrystalline diamond (NCD) [6], [7] to micrometers, < 100 μm [8], [9] in polycrystalline diamond (PCD). Typical deposition conditions involve the use of a gas mixture containing a small quantity of hydrocarbon gas, e.g., 1% CH4 in excess of hydrogen, which normally affords PCD films. Nanocrystalline diamond films have been produced in various ways: 1) by increasing the CH4 percentage from 10% to 20% in the standard CH4–H2 mixture [10], 2) by biasing the untreated Si substrate on which the film is deposited during the growth in MWPECVD [11], [12] or 3) by replacing H2 with inert gases [6], [13].

Specifically, in Ar-rich (> 95%) Ar–H2–CH4 gas mixtures, the grain size significantly decreases, becoming of the order of nanometers. As a consequence, the morphology of the film itself changes drastically at the nanometer scale. UNCD and NCD films obtained in this way result smoother than PCD films. In Ar-rich Ar–H2–CH4 gas mixtures, the presence of H2 was found to confer stability to the microwave plasma. On the other hand, too high H2 percentages would lead to materials showing a high surface roughness. In the present work, diamond films were grown on silicon substrates screening different quantities of H2 from 15% down to 6.3%. The plasma phase and plasma–substrate interface were investigated by in-situ optical emission spectroscopy (OES) and pyrometric interferometry (PI) to determine the behavior of emitting species and the deposition rates, respectively. The obtained films were finally characterized by Raman spectroscopy, atomic force microscopy (AFM) and scanning electron microscopy (SEM) techniques. The films were grown on an initial buffer layer (BL) about 500 nm thick, which allowed the production of uniformly deposited samples, in contrast with the coatings obtained without BL that were non continuous and/or covered with pinholes. In particular, for H2 percentages between 6.3 and 10%, the presence of a buffer BL afforded films having nanocrystalline structure and morphology, generally showing low root-mean-squared (rms) roughness values.

Section snippets

Experimental

Diamond films were deposited on 2.6 × 2.6 cm2 square pieces of polished p-doped (B) silicon (100) substrates by MWPECVD technique, in a home-made cylindrical stainless steel 2.45 GHz Astex-type reactor. Prior to the deposition process, the Si substrates were ultrasonically seeded for 1 h with an ethanol suspension of 40–60 μm diamond powder. The pre-treated substrates were placed on a molybdenum holder set on a graphite susceptor. The susceptor temperature was controlled by an external radiative

Results

Four separate depositions with BL at different hydrogen percentages (6.3%, 8%, 10% and 15%) in the gas composition were performed. Having fixed the methane percentage in the mixture to 1%, the argon percentage was adjusted as a consequence of the hydrogen variation. For 6.3% of H2 two coatings were also obtained without BL and were used for comparison purposes.

Fig. 3 shows the evolution of the C2/Hα, CH/C3 and Hα/Ar OES intensity ratios as a function of the vertical distance from the

Discussion

The percentage of H2 in the gas mixtures inevitably affects the amount of the generated carbon species and hydrogen atoms, as seen in Fig. 3, Fig. 4. Although the optical emission spectroscopy cannot be relied as quantitative technique for the limits well described in detail in Refs. [18], [19], still the study is helpful for monitoring the NCD film deposition reactor parameters. The plasma chemistry of CH4 in Ar-rich mixture is very complex and involves both dissociation (Eq. (6)) and

Conclusions

The buffer layer is an expedient that allows the production of uniformly deposited samples having small roughness in contrast with coatings without buffer layer being non continuous and/or covered with interfacial or surface pinholes. The preliminary results of roughness improvement, in the film obtained at 10% of H2 (promoting stable plasma conditions) by decreasing the working pressure (140 mbar, Fig. 6B) during the deposition process, evidence the beneficial effect of the buffer layer.

Acknowledgments

This work was carried out in the frame of “Progetto Strategico” ATS PS_136 of Regione Puglia (Italy).

References (21)

  • A. Gicquel et al.

    Curr. Appl. Phys.

    (2001)
  • J. Birrell et al.

    Diamond Relat. Mater.

    (2005)
  • G. Cicala et al.

    Diamond Relat. Mater.

    (2005)
  • G. Cicala et al.

    Surf. Coat. Technol.

    (2010)
  • M.A. Nitti et al.

    Diamond Relat. Mater.

    (2011)
  • T. Sharda et al.

    Diamond Relat. Mater.

    (2000)
  • O.J.L. Fox et al.

    Diamond Relat. Mater.

    (2009)
  • G. Davies
  • P.K. Bachmann
  • S. Jiao et al.

    J. Appl. Phys.

    (2001)
There are more references available in the full text version of this article.

Cited by (12)

  • Three elements for the preparation of vertical graphene by RF-PECVD method

    2021, FlatChem
    Citation Excerpt :

    The plasma is attached or grown on the sample substrate to form a vertical graphene film perpendicular to the surface of the sample. Figure 2 shows the preparation of vertical graphene, initially adopted PECVD technology, which developed to three different preparation modes, direct current (DC) plasma technology, microwave (MW) plasma technology, and RF plasma technology [38-45]. Direct Current Plasma Enhanced Chemical Vapor Deposition (DC-PECVD) is the first technology applied to PECVD to prepare vertical graphene.

  • Study on the effect of Ar-containing work gas on the microstructure and tribological behavior of nanocrystalline diamond coatings

    2021, Tribology International
    Citation Excerpt :

    The microstructure and morphology of NCD can be controlled by many growth dynamic factors, such as pre-seeding density and nucleation density [17–19], substrate and filament temperature [20,21], and work gas pressure [22–24]. With Ar plasma assistance, it is assumed that the active precursor group is the C2 dimer, which lowers the NCD growth activity energy [12–15,25–30]. In contrast, it is also thought that the active precursor is still CH3, even with plasma assistance, because the sticking coefficient of C2 is much smaller than that of CH3, and the measured activity energy is nearly identical [1,20,30].

  • Emittance evaluation of electron beams produced by photoemission

    2020, Nuclear Instruments and Methods in Physics Research, Section B: Beam Interactions with Materials and Atoms
    Citation Excerpt :

    In a NEA material, the vacuum level is shifted beneath of conduction band minimum and therefore the electrons injected in the conduction band are emitted directly into the vacuum, improving the photoemission. The production of diamond films is based on chemical vapour deposition (CVD) techniques [17–20], which work at high substrate temperatures (≥600 °C). In the last years, new techniques have been developed in diamond technology, to overcome the limit of the high deposition temperature and to allow different possibilities in terms of materials utilized as carbon sources.

  • Growth of nanocrystalline diamond by dual radio frequency inductively coupled plasma jet CVD

    2017, Diamond and Related Materials
    Citation Excerpt :

    It can be seen that the major optical emission lines are from atomic Ar (the range of 696.5 to 850 nm), H atom (Hα @656.3 nm, Hβ @486.1 nm) as well as hydrocarbon related radicals (CH @431.5 nm, CH+ @422.1 and 417.1 nm, C2 @Swan band 474.7, 516.1, 563.6 and 619.2 nm). Similar OES can be seen in other experimental cases [26]. The intensity ratio of Hβ, CH and C2 to Hα versus methane concentration is plotted in Fig. 3.

View all citing articles on Scopus
View full text