(Color online) Normalized DLTS spectra taken on samples with (1) high (electron irradiated, dotted line), (2) standard (as-grown (commercial), solid line), (3) low (as-grown (Kyoto-University), dashed line), and (4) very low (oxidized, dash-dotted line) initial -concentration, which (a) have not been annealed and (b) have been annealed at 1850 °C for 15 min. The annealed samples show nearly identical -signals, which are even higher than that of sample (1) (Fig. 1(a)). The spectra depict the sine-correlation b 1 normalized by net doping and capacitance under reverse bias at room temperature with a time window of . The measurements were taken with reverse bias of , pulse bias of , and filling pulse time of .
(Color online) (a) - and (b) -concentration of set (1) to (4) vs annealing temperature. The very left measurement points show the by-orders-of-magnitude varying initial concentration of the reference samples, which have not been annealed. In the case of the oxidized samples, the concentrations are an upper limit, as they lie below the detection limit (open circles).
(Color online) (a) - and (b) -concentration vs reciprocal annealing temperature. From the slope of the Arrhenius plot for the sample sets (3) and (4), the formation energy is determined after Eq. (3) as (a) and (b) .
(Color online) Mapping of the μ-PCD lifetime for non-annealed samples (first row) and samples which have been annealed at 1500 °C (second row) and at 1850 °C (third row), respectively, of each sample set. τ is relatively uniformly distributed inside each sample and depends on the respective -concentration.
(Color online) Lifetime obtained by μ-PCD as a function of the annealing temperature. The lifetime is decreasing as the concentration of is reaching values larger than.
Overview of the investigated sample sets with an epilayer thickness , a net nitrogen doping , and an initial -concentration .
Article metrics loading...
Full text loading...