Elsevier

Carbon

Volume 41, Issue 6, 2003, Pages 1301-1308
Carbon

Residual stresses and crystalline quality of heavily boron-doped diamond films analysed by micro-Raman spectroscopy and X-ray diffraction

https://doi.org/10.1016/S0008-6223(03)00071-XGet rights and content

Abstract

X-ray diffraction analysis and micro-Raman spectroscopy measurements have been used for stress studies on HFCVD diamond films with different levels of boron doping. The boron incorporation in the film varied in the range 1018–1021 boron/cm3. The grain size, obtained from SEM images, showed grains with 2–4-μm average size, which decreases when the doping level increases. The thickness of the films obtained by SEM cross-section view decreased from 8 to 5 μm as the doping level increased from 0 (undoped film) to 1021 boron/cm3. The total residual stress was determined by measuring, for each sample, the (331) diamond Bragg diffraction peak for Ψ-values ranging from −60° to +60°, and applying the sin2 ψ method. For the micro-Raman spectroscopy the spectral analysis performed on each sample allowed the determination of the residual stress, from the diamond Raman peak shifts, and also the diamond purity, which decreases from 99 to 75% as the doping level increases. The type and magnitude of the residual stress obtained from X-ray and micro-Raman measurements agreed well only for undoped film, disagreeing when the doping level increased. We attributed this discrepancy to the domain size characteristic of each technique.

Introduction

The study of residual stress in synthetic diamond films has been subject of intense research in the last few years [1], [2], [3]. Diamond grown on different substrates presents various singular properties and, due to its wide applications, many efforts have been made to measure the stresses and to explain their origin [4], [5]. Particularly, semiconductor diamond research evolution has shown promising results with a doping process during CVD growth. Highly boron-doped diamond electrodes are important for electrochemical studies, offering high sensitivity, good precision and high stability when compared with vitreous carbon. Its more remarkable property is a wide working potential window [6]. In this sense, the introduction of a large concentration of boron atoms in CVD diamond may promote significant changes in its lattice parameter and may introduce defects increasing the impurities in the grain boundaries.

Residual stress, mainly for undoped CVD diamond films, has been reported by many authors in previous studies using different techniques, such as the substrate curvature technique [1], X-ray diffraction [3], [7] and Raman spectroscopy [2]. Particularly, Raman spectroscopy is a simple method that evaluates the residual stress through the diamond Raman peak shifts and detects a local strain inside the grains. However, when micro-Raman is used, quantitative evaluation may be difficult due to the domain size effect [8] observed by the line shape and also, multiple peaks may appear attributed to peak splitting due to degeneration of the optical phonons and the presence of inhomogeneous micro stresses. In order to measure the residual stress by X-ray diffraction, the sin2 ψ technique is usually applied. This method allows determination of the residual stress averaged over a larger sample area, and due to the transparency of diamond to X-rays, through the whole film depth. The results generally show that the global intrinsic stress depends on the methane fraction, deposition temperature [1] and crystallographic orientation [9]. However, understanding of residual stress formation in such films is incomplete and there is significant disagreement among various researchers as to the type and magnitude of the stress, even for similar deposition conditions. This divergence may be attributed to the stress measurement techniques, which generally have intrinsic limitations that should be considered. This behavior was extensively discussed in previous work [10] for undoped diamond films on a silicon substrate as a function of the film thickness.

For boron doped diamond films, adherence becomes a limitation for producing highly doped electrodes for using in electrochemical analysis in corrosive solutions. Knowledge and control of the residual stress in such films are important in this study. In the case of boron-doped diamond films, boron incorporation on substitutional or interstitial sites can produce stresses according to the doping level. Recently we have also demonstrated the use of X-ray diffraction technique for doped diamond films [11]. In this paper we continue this study by comparing the two techniques for stress evaluation. Stresses on boron-doped diamond films are analyzed from X-ray diffraction measurements by using the sin2 ψ method and micro-Raman spectroscopy, at different levels of boron doping; to our knowledge there are no reports on stress studies which compare these two methods in such films. The morphology, thickness and film quality were also analyzed by SEM and Raman spectroscopy. A large dependence of the level of tungsten contamination in the film with doping level was observed with synchrotron X-ray radiation. The presence of tungsten carbide in the film, evaluated as up to 1% in volume for the most doped films, certainly affects film structure and consequently, the stress results.

Section snippets

Experimental

The films were deposited on silicon substrate after seeding pre-treatment [12] in a hot filament-assisted CVD reactor by using a gas flow rate of 0.5 sccm of methane, 89.5 sccm of pure hydrogen and 10 sccm of hydrogen passed through a bubbler containing B2O3 dissolved in methanol. The bubbler temperature and pressure were kept constant at 300 K and 2.0×102 Pa, respectively, in order to guarantee the same basic gas mixture in the reactor. The boron to carbon (B/C) ratio in the gas mixture was

Scanning electron microscopy and Raman spectroscopy

The surface analysis made by SEM for boron-doped diamond films showed a small variation of surface morphology in comparison to undoped diamond. The grains are faceted with symmetrical and smooth faces with uniform texture that show a surface morphology with predominant (111) orientation. For all doping levels studied the films showed grains of 2–4-μm average size, which decreases when the doping level increases. The thickness of the films shown in an SEM cross-section view also decreased from 8

Discussion and conclusion

The comparison between the residual stress obtained from the X-ray measurements and micro-Raman analysis is plotted in Fig. 7. The results show agreement only for the undoped film. Raman analyses showed a drastic increase of tensile residual stress as a function of boron incorporation while for X-ray diffraction the stress values become more and more compressive with the increase of boron incorporation in the diamond lattice.

We cannot establish here the fundamental reasons for the discrepancy

Acknowledgements

We are very grateful to CNPq (process No. 81208/97-4) and FAPESP (process No. 95/6219-4) for financial support. We are also grateful to M.L. Brison for SEM analysis and W. Brazolin for technical support.

References (27)

  • H. Windischmann et al.

    Diam Relat Mat

    (1995)
  • C.T. Kuo et al.

    Thin Solid Films

    (1996)
  • M. Hempel et al.

    Diam Rel Mat

    (1999)
  • N.G. Ferreira et al.

    Diam Rel Mat

    (2001)
  • R.C. Mendes de Barros et al.

    Diam Rel Mat

    (1996)
  • J. Cifre et al.

    Diam Rel Mat

    (1994)
  • F. Brunet et al.

    Thin Solid Films

    (1998)
  • N.G. Ferreira et al.

    Diam Rel Mat

    (2002)
  • R.J. Zhang et al.

    Diam Rel Mat

    (1996)
  • B.V. Spitsyn et al.

    J Cryst Growth

    (1981)
  • H. Windischmann et al.

    J Appl Phys

    (1991)
  • D. Rats et al.

    J Appl Phys

    (1995)
  • Y. Von Kaenel et al.

    J Appl Phys

    (1997)
  • Cited by (107)

    View all citing articles on Scopus
    View full text