We demonstrate that in situ post-growth annealing of GaMnAs layers under As capping is adequate for achieving high Curie temperatures (T C ) in a similar way as ex situ annealing in air or in N 2 atmosphere practiced earlier. Thus, the first efforts give an increase of T C from 68 K to 145 K after 2 hours annealing at 180 °C. These data, in combination with lattice parameter determinations and photoemission results, show that the As capping acts as an efficient sink for diffusing Mn interstitials. In recent years, considerable progress has been done both in understanding the material properties, and improving the quality of low-temperature molecular beam epitaxy (LT MBE) grown GaMnAs films [2][3][4][5] . This has led to an increase of T c in GaMnAs from the previously "established" limit 6 of 110 K to 160 -170 K, as is now reported by several research groups 4,7,8 . The increase of T c by 50 -60 K was possible due to the identification and understanding of the role of structural defects, which compensate the Mn acceptors in GaMnAs. Since GaMnAs can only be grown by LT MBE, it hosts all the defects typical to LT GaAs 9-11 . At the MBE growth conditions suitable for GaMnAs deposition the prevailing structural defect is the As antisite (As Ga ). It is well known 10-12 that the maximum concentration of As Ga is close to 0.5%. Similar or even higher concentration of As Ga is expected in GaMnAs 13, 14 . In addition there is one defect specific to GaMnAs, namely Mn in interstitial positions (Mn I ). The importance of this defect was first analyzed theoretically 15 , and then Mn I was identified experimentally 16 . The negative impact of Mn I on the ferromagnetic properties of GaMnAs is due to two effects: Mn I acts as a double donor, and thus tends to compensate the p-doping In this letter we analyze the magnetic, structural and electronic properties of (Ga,Mn)As that is annealed under an As capping layer. Since the As capping protects the surface from 3 oxidation, the annealing can be carried out in air as well as in vacuum. We show that there is no essential difference in the annealing efficiency between the two environments. The behavior of the As cap during annealing in vacuum was monitored by reflection high energy electron diffraction (RHEED). The annealing temperature in vacuum was measured with an infrared pyrometer, and in air by means of a thermocouple in contact with the Mo-block holding the In-glued (Ga,Mn)As samples.We have grown a set of Ga 0.94 Mn 0.06 As layers with thicknesses in the range of 100 Å to 1000 Å, keeping the MBE growth procedure the same for all samples. Before sample growth, the Ga and Mn fluxes were carefully calibrated using RHEED intensity oscillations from a separate test sample. The GaMnAs samples were grown on semi-insulating GaAs(100) epiready substrates. Samples intended for photoemission were grown on n-type (Te doped) substrates. Each growth started with deposition of a 1000 Å thick high temperature (HT)GaAs buffer, and all samples were grown at the maximum substrate temperat...

The influence of annealing parameters -temperature (Ta) and time (ta) -on the magnetic properties of As-capped (Ga,Mn)As epitaxial thin films have been investigated. The dependence of the transition temperature (TC) on ta marks out two regions. The TC peak behavior, characteristic of the first region, is more pronounced for thick samples, while for the second ('saturated') region the effect of ta is more pronounced for thin samples. A right choice of the passivation medium, growth conditions along with optimal annealing parameters routinely yield TC-values of ∼ 150 K and above, regardless of the thickness of the epilayers.

In the quest for ferromagnetic semiconductors suitable for future spintronics, the (GaMn)As system remains a potential candidate material, particularly after the discovery of possibilities to raise its Curie temperature (T C ) by post-growth annealing [1]. Some important experimental observations concerning the annealing process are well established. The surface conditions play an important role -until now all successful annealing treatments have been performed in air or in N 2 atmosphere, while annealing in vacuum or with a protecting GaAs capping is found to be inefficient. The treatment also becomes gradually less efficient with increasing thickness of the (GaMn)As layer. While it is generally agreed that the annealing effects are due to diffusion and surface passivation of Mn interstitials, detailed understanding of the process is still lacking. In a recent Letter [2] Edmonds et al. assumed that Mn diffusion is the rate-limiting factor for annealing-induced changes. From measurements of layer resistivity as function of annealing time these authors then determined the activation energy for Mn diffusion. However, there is an obvious alternative rate limiting mechanism, namely the trapping efficiency of the diffusing Mn interstitials (as suggested by the mentioned surface sensitivity). In this Comment we demonstrate that this mechanism is indeed the active one in the present case, and this invalidates the results in Ref. 2.