next up previous contents
Next: Electronic structure of defects Up: Impurities and defects in Previous: p-type diamond

n-type diamond

Potential n-type dopants in diamond may be found in the group V elements, as substitutional donors. However, n-type doping appears to be extremely difficult. Nitrogen is abundant in natural types Ib and Ia diamonds, which can explain the negative formation energy of -3.4 eV calculated by ab initio theory [23]. The activation energy of the nitrogen donors is measured to be 1.7 eV in type Ib diamond, and 4 eV in type Ia diamond. These deep levels make therefore the nitrogen doped diamond insulating at room temperature. Kalish et. al. [61] introduced nitrogen in diamond by implantation, followed by annealing at high temperature. Almost half of the nitrogen implanted was found to be located at subtitutional sites, as found in natural diamond, giving rise to deep levels in the band gap. Thus, nitrogen doped diamond is useless in electronic devices, even at high temperature.

The next potential dopant in the type V elements is phosphorus. Kajihara et. al. [23] calculated that phosphorus on a substitutional site is a shallow donor, with an energy level of 0.2 eV. However its high formation energy (10.4 eV [23]) leads to a small solubility, which makes the phosphorus doping by in-diffusion ineffective. Indeed, Borst et. al. [62], for instance, failed in doping diamond during the CVD process, and the activation energy as well as the phosphorus concentration in the sample could not be measured. Very recently, Koizumi et. al. [63] introduced phosphorus during the growth process by CVD and obtained a donor level at 0.5 eV below the bottom of the conduction band. However the mobility measured was found to be very low, namely between 30 and 180 cm2/V s.

The reliable way to dope diamond with phosphorus should therefore be by ion implantation. Prins [33] implanted phosphorus at low dose to reduce as much as possible the concentation of defects created during the ion implantation process. This was followed by low temperature annealing to drive the dopant toward subsitutional sites. Various activation energies were measured, at around 0.2 eV, with n-type conductivity deduced only from hot probe measurements. However it should be mentioned that no Hall measurements were reported in this work, and that a control experiment (see above) was not attempted to accredit the results.

Ran [64] also attempted to introduce phosphorus into diamond by implantation, followed by annealing In this experiment, phosphorus atoms were found up to a depth of 0.25 $\mu$, and the phosphorus density obtained was $\sim 10 ^{19}$ P/cm3. To ovoid the formation of too many defects during the implantation, a high temperature was applied during the introduction of the dopants (1070 K), and in the post-implantation annealing (1600 K). However, it was found that the phosphorus atoms were not electrically activated at the end of the process. It was speculated that too many defects have been created during the implantation, because of the large size of the phosphorus atom. Thus, although half of the dopant atoms are located on substitutional sites, the rest are traped by defects, forming complexes, like phosphorus-vacancy complex, that may act as deep acceptors centers [65]. The later may therefore neutralize possible active phosphorus donors.

Lithium and sodium donor impurities have also been the subjects of theoretical and experimental studies. Since both of them have only one atom in the external shell, they are expected to act as donors when located at interstitial sites. The energy level calculated by Kajihara et. al. [23] was 0.3 eV below the bottom of the conduction band for Na, and 0.1 eV for Li, with respective formation energies of 15.3 eV and 5.5 eV. The high formation energies of these two dopants is associated with a low solubility, which should lead to difficulties in incorporating these dopants into diamond by diffusion or during growth.

Many attemps were made in the past to doped diamond with lithium, either by diffusion, by incorporation during CVD growth or by implantation [66,67,68]. In the diffusion [66] and growth [67] experiments, the results obtained were ambiguous and did not lead to any conclusion as to what extent lithium is a reliable useful donor in diamond.

During the implantation process, the lithium atoms are expected to create relatively few defects, due to their light mass and small dimension. Hot lithium implantation into diamond at 850 C, followed by annealing at 900 C, was performed by Job et. al. [69]. An activation energy of $\sim$ 0.2 eV was measured, however with a very high resistivity. Buckley-Golder [70] introduced lithium into a type IIb diamond by implantationat at high temperatures. Due to the appearance of a p-n junction behavior in the samples, it was speculated that lithium acts as a donor, the acceptor part being associated with the boron atoms existing in type IIb diamond. However the effect of the defects created during the implantation was not investigated, although their presence induces energy levels in the band gap. Moreover, it was not clearly determine wether lithium create shallow donor level. Prawer et. al. [68] also implanted lithium at different doses and at low temperatures, followed by annealing at high temperature. Lithium was indeed found inside the sample, however its resistivity was too high to identify the type of conductivity encountered.

Sodium was also implanted in diamond by Prawer et. al. [68]. However, as for lithium, no conclusive indication of n-type conductivity could be extracted from the experimental results. For both these cases, hopping conductivity was found.


next up previous contents
Next: Electronic structure of defects Up: Impurities and defects in Previous: p-type diamond
David Saada
2000-06-22