Boron delta doping of diamond

In an attempt to combine in the same single crystal diamond device a high carrier density at room temperature and a high mobility, we have undertaken to grow diamond epilayers delta-doped with boron. Thanks to a modified reactor and gas feeding system and to original growth and in situ etch-back strategies, heavily doped epilayers thinner than 2 nm were obtained. However, not any of the expected confinement-induced carrier delocalization and resulting mobility enhancement effects could be observed.

One of the main issues limiting the impact of diamond in electronics is the large ionisation energy of the acceptor level (0.38 eV for B), which entails a very low carrier concentration at room temperature. Metallic diamond ([B]> 5 1020/cm3, i.e. 0.3 at.‰) does provide a high temperature-independent carrier density, but the mobility in the valence band is low, on the order of 1 to 5 cm2/Vs, because of the high density of ionised impurities. Combining a high carrier density with a high mobility would put diamond on the roadmap of high frequency power devices. One possibility is to confine the free holes of the metal into a quantum well thin enough (the “delta-doped” layer) to induce carrier delocalisation away from the ionised impurities, into the nearby undoped diamond layers where the free hole mobility lies above 1000 cm2/Vs at room temperature. This idea has been proposed 20 years ago, but despite the absence of diffusion of dopants in diamond, its efficient application to this material has remained quite elusive. The challenges delta-doping poses to epitaxial growth are not restricted to the nanometric thickness and very high doping level, but involve also the sharpness or electronic quality of the interfaces with the surrounding low doped epilayers ([B]< 5 1017/cm3). These stringent requirements challenge moreover physico-chemical characterisation methods such as Secondary Ion Mass Spectroscopy (SIMS) or Transmission Electron Microscopy (TEM). In collaboration with the deltadiam ANR consortium, the NIMS at Tsukuba (Japan), the SIMS facility of GEMaC, and TEM experts of Univ. Cadiz (Spain), we decided to tackle this issue.

Fig. 1 : (top) STEM-HAADF images of two delta-layers (darker areas) labeled  and  of respective thicknesses 2 and 3 nm ; (bottom) SIMS profiles of the boron content for 2 analysing ions and energies.

We first modified a home-made microwave-enhanced chemical vapor deposition reactor to allow ultrarapid switching of the gas mixtures. This enabled us to grow delta-layers without turning off the plasma, an original asset for the chemical purity of the interface. In addition, we developed in situ etch-back processes in H2 or O2/H2 atmospheres to control even better the thickness and the second interface1. Finally, we introduced in situ monitoring of the growth and etch rates by spectroscopic ellipsometry. As shown in Fig. 1 (bottom), SIMS profiles of such epilayers were broadened by strong ion-mixing effects, which were evaluated on isotopically substituted delta-layers. The derived minimum thicknesses in the 1-3 nm range were checked against those obtained from high angle annular dark field (HAADF) in a scanning TEM2 (Fig.1), and from the simulation of ex situ ellipsometry spectra. As shown in Fig. 2 (top), Hall effect and resistivity measured on metallic samples led to similar thicknesses. Temperature-dependent mobility measurements of delta-layers with thicknesses decreasing below 2 nm did not show any sign of confinement-enhanced mobility3. They yielded a value of 3.6 cm2/Vs independent of the thickness and similar to that of the bulk. As shown in Fig. 2 (bottom), this value was also found to be significantly lower than calculated for 3D ionised impurity scattering.

Fig. 2 : (top) sheet iso-density lines for various metallic delta-doped epilayers, allowing to deduce maximum thicknesses from Hall effect ; (bottom) experimental mobility values compared to that calculated for ionised impurity scattering in the bulk at the same boron doping level.

1 A. Fiori, F. Jomard, T.N. Tran Thi, G. Chicot, F. Omnès, E. Gheeraert, E. Bustarret, “In situ etching-back processes for a sharper top interface in boron delta-doped diamond structures”, Diamond & Rel. Mat. 24, 175 (2012). 2 D. Araujo, M.P. Alegre, J.C. Piñero, A. Fiori, E. Bustarret, F. Jomard, « Boron concentration profiling by HAADF-STEM in homoepitaxial -doped diamond layers », Appl. Phys. Lett. 103, 042104 (2013). 3 G. Chicot, T.N. Tran Thi, A. Fiori, F. Jomard, E. Gheeraert, E. Bustarret, J. Pernot, “Hole transport in boron delta-doped diamond”, Appl. Phys. Lett. 101, 162101 (2012).

Fig. 1 : (top) STEM-HAADF images of two delta-layers (darker areas) labeled  and  of respective thicknesses 2 and 3 nm ; (bottom) SIMS profiles of the boron content for 2 analysing ions and energies.
© Institut Néel 2012 l Webdesign l Powered by spip l Last update dimanche 2 février 2020 l