EL ANÁLISIS DE LA RAZONABILIDAD DE LAS BARRERAS BUROCRÁTICAS

As discussed in [14] and [62] hydrogen is a crucial component of the CVD source gas for diamond growth. The prevalence of hydrogen in the growth environment invites

consideration of the incorporation of this impurity into the diamond. Hydrogen has been shown to have an affect on the electronic performance of diamond, for

example, by compensating the substitutional boron acceptor [63]. Consequently, the trapping or release of hydrogen by point or extended defects is an important

topic for research.

Hydrogen is observed in CVD diamond at the surface and crystal grain bound- aries [64]. For example, the H1 EPR-active centre [65, 66] which involves a single

hydrogen atom is present only at grain boundaries. Hydrogen can also be incorpo- rated into the crystal bulk through growth errors, which are grown over resulting

in a point defect [67].

Hydrogen-related defects are also observed optically; a review is given by Crud-

dace [68]. One prominent hydrogen-related optical transition is the 3107 cm−1 _ab-

sorption line which is commonly observed in natural and HPHT treated diamonds.

The 3107 cm−1 _{line is thought to originate from a carbon-hydrogen stretch mode}

[69, 70]. The 3123 cm−1 _{absorption line, that will be discussed in Chapter 6 in-}

volves a hydrogen and a carbon atom but has no isotopic shift with15_{N enrichment}

[71].

Two routinely observed hydrogen containing EPR-active defects are reviewed

here.

2.4.1

The vacancy-hydrogen complex

The VnH− defect is commonly observed in CVD diamond [72, 73]. The defect is

labelled here with an unspecified number of vacancies ‘n’ as there are currently two

models proposed for this defect. The first, initially proposed by Glover et al. [73], involves a hydrogen atom bonded to one of four carbon neighbours surrounding

Chapter 2. Literature review

a vacancy (n=1), see Figure 10.2(b). However, density functional theory (DFT) calculations by Shaw et al. [72] predict that the hydrogen hyperfine interaction resulting from such a construction would be significantly larger than that observed

by the experiment. They instead proposed that the experimentally observed hy- perfine interaction is consistent with a dynamic defect. This second model suggests

that the hydrogen atom is tunnelling between three equivalent dangling bonds at one end of a di-vacancy (n=2), see Figure 2.5(b).

EPR spectra recorded at temperatures in the range 4 K to 300 K have shown no evidence of motional averaging on an EPR time scale [73]. However, the tunneling

between orbitals may have such a low energy barrier that it is unaffected by this temperature change. This means that we cannot differentiate between the models

based on the available EPR data. The VnH− defect has also been observed in

deuterium enriched samples (VnD−), but no splitting from the deuterium hyperfine

interaction was resolved [68].

2.4.2

The nitrogen-vacancy-hydrogen complex

EPR spectra resulting from the negative nitrogen-vacancy-hydrogen NVH− _defect

were first observed by Hunt, [50] and subsequently by Gloveret al. [74]. The NVH−

defect is routinely observed in SC-CVD diamond, but has not been reported in natural or HPHT synthetic diamond. Cruddace [68] showed that the defect can

be responsible for a significant fraction of the nitrogen in SC-CVD diamond (up to 10%). These measurements were made on samples with N0

S concentrations of

100–1000 ppb.

EPR measurements show that the NVH− _{defect has an effective C}

3v symme-

try [74]. Theoretical DFT calculations by Kerridge et al. [75] and Shaw et al.

[72] have predicted that the hydrogen would tunnel between three equivalent po-

sitions, where it bonds to one of the three carbon atoms surrounding the vacancy. The static configuration of the defect would have C1h symmetry. Edmonds [76]

has recently shown that the EPR observations are consistent with this dynamic model and that the observed EPR spectrum results from motional averaging of

the configuration with C1h symmetry.

Chapter 2. Literature review Ca C_b C_c C_d [110] C_b C_e C_f C_{a [110]} [001] H C_H V (a) C_a C_b C_c C_d [110] C_b C_e C_f C_{a [110]} [001] V V H (b)

Figure 2-5: Cartoon depictions of the VnH− defect in diamond projected in the

(1¯10) plane. The larger circles represent those carbon atoms in the foreground, the smaller dotted circles represent those in the plane behind when viewed in this direction. (a) the V1H model proposed by Glover et al. [73] and (b) the V2H

model proposed by Shaw et al. [72]. The neighbouring carbon atoms are labelled in groups with equivalent positions. The broken squares represent vacancies and the shaded circle a hydrogen atom. In the V2H model the hydrogen atom is

tunneling between the dangling bonds from the three Cb atoms, as indicated by

Chapter 2. Literature review

rules that forbid it [77]. Goss et al. [77] made predictions for the positions of the local vibrational mode, originating from a C-H stretch mode, of the NVH defect in both negative and neutral charge states at ∼2827 cm−1 _and ∼_{2679 cm}−1

respectively. Caution should be exercised here since these predictions are at best ‘semi-quantitative’ [77].

Cruddace [68] showed that the 3123 cm−1 _{absorption line resulted from a defect}

with C1h symmetry and correlated its intensity with the concentration of NVH−

over four orders of magnitude. It was subsequently suggested by Meng et al. [78] that the 3124 cm−1 _line1 _{may be associated with the NVH}− _defect.