We report on a comprehensive combined experimental and theoretical study of Curie temperature trends in ͑Ga,Mn͒As ferromagnetic semiconductors. Broad agreement between theoretical expectations and measured data allows us to conclude that T c in high-quality metallic samples increases linearly with the number of uncompensated local moments on Mn Ga acceptors, with no sign of saturation. Room temperature ferromagnetism is expected for a 10% concentration of these local moments. Our magnetotransport and magnetization data are consistent with the picture in which Mn impurities incorporated during growth at interstitial Mn I positions act as double-donors and compensate neighboring Mn Ga local moments because of strong nearneighbor Mn Ga u Mn I antiferromagnetic coupling. These defects can be efficiently removed by post-growth annealing. Our analysis suggests that there is no fundamental obstacle to substitutional Mn Ga doping in high-quality materials beyond our current maximum level of 6.8%, although this achievement will require further advances in growth condition control. Modest charge compensation does not limit the maximum Curie temperature possible in ferromagnetic semiconductors based on ͑Ga,Mn͒As.

Edmonds et al. Reply: (Ga,Mn)As is a test ground of choice in semiconductor-based spintronics research. This is due, in particular, to the development of annealing procedures that substantially improve its electric and magnetic properties. However, the quality of (Ga,Mn)As has not been fully optimized yet; thus, understanding the exact nature of defects and their annealing and passivation is highly interesting. The Comment by Adell et al. [1] represents an attempt in this direction.Adell et al. study the influence of annealing of (Ga,Mn)As covered by a capping layer of amorphous As on the Curie temperature T C . They confirm the main conclusion of our paper [2], namely, that the removal of defects by annealing of (Ga,Mn)As occurs in two steps: (i) the out diffusion of interstitial Mn (Mn:I) toward the surface, followed by (ii) electrical passivation of Mn:I donors by the formation of, e.g., an oxide (which increases the hole concentration and the Curie temperature). In [1], Mn:I diffuses towards the (Ga,Mn)As/As interface, where it is passivated by formation of MnAs.Comparing Ref.[1] with Ref.[2] it is noted that (a) during the first two hours the kinetics of annealing is faster in [1] (however, not ''by 1-2 orders of magnitude'' as is claimed by Adell et al.); (b) time required to reach the maximum T C in [1] is independent of the thickness of the samples, in contrast to Ref.[2]; (c) in [1] the initial rise of T C is followed by a degradation of sample, while we find that annealing of uncapped samples leads to a continuous improvement, even after much longer times, and higher T C .From the point (a) Adell et al. suggest that in [2] the process limiting the kinetics of annealing is not the bulk diffusion of Mn:I but the formation of a protective oxide layer at the surface. Although surface accumulation of Mn oxide certainly occurs in air-annealed samples, as demonstrated by the Auger spectra in Fig. 2 of [2], we see no evidence that the passivation rate decreases as the oxide thickness increases. Such an effect would break the (1=L 2 )-scaling of the diffusion rate, which is clearly observed in our study over a 2 orders of magnitude change of L 2 . We agree that in principle a reduced passivation rate will reduce the annealing efficiency, however the results of Ref.[1] are too incomplete to be conclusive.Point (b) is ascribed in [1] to the formation of a reacted MnAs layer that hinders further passivation. Therefore, the ''optimum'' anneal time observed in [1] corresponds to the time required to form a reacted layer, rather than the time required for complete out diffusion of interstitials.The degradation induced by longer annealings, point (c) above, is unexplained, and is usually only observed on annealing at much higher temperatures [3]. Since we do not observe this effect, it should be due to the presence of the As capping layer. One may speculate that this is caused by the in diffusion of As and formation of As-induced donors, e.g., As antisites, facilitated by the low position of the Fermi level. Independe...

We study the dc transport properties of (Ga,Mn)As diluted magnetic semiconductors with Mn concentration varying from 1.5% to 8%. Both diagonal and Hall components of the conductivity tensor are strongly sensitive to the magnetic state of these semiconductors. Transport data obtained at low temperatures are discussed theoretically within a model of band-hole quasiparticles with a finite spectral width due to elastic scattering from Mn and compensating defects. The theoretical results are in good agreement with measured anomalous Hall effect and anisotropic longitudinal magnetoresistance data. This quantitative understanding of dc magneto-transport effects in (Ga,Mn)As is unparalleled in itinerant ferromagnetic systems.

We report on measurements of the hole density, ferromagnetic transition temperature TC, and magnetization in a series of as-grown and annealed (Ga,Mn)As samples. Estimating the fraction of incorporated Mn occupying interstitial and substitutional sites allows a direct comparison of the predictions of mean field theory with experiment, and a determination of the magnetic moment per substitutional Mn. The saturation of TC at high Mn concentration is consistent with the mean field prediction. The estimated magnetic moment per Mn is close to the expected 5μB for all samples studied if an antiferromagnetic coupling between interstitial and substitutional Mn is assumed.

We have performed a systematic investigation of magnetotransport of a series of asgrown and annealed Ga 1-x Mn x As samples with 0.011 ≤ x ≤ 0.09. We find that the anisotropic magnetoresistance (AMR) generally decreases with increasing magnetic anisotropy, with increasing Mn concentration and on low temperature annealing. We show that the uniaxial magnetic anisotropy can be clearly observed from AMR for the samples with x ≥ 0.02. This becomes the dominant anisotropy at elevated temperatures, and is shown to rotate by 90 o on annealing. We find that the in-plane longitudinal resistivity depends not only on the relative angle between magnetization and current direction, but also on the relative angle between magnetization and the main crystalline axes. The latter term becomes much smaller after low temperature annealing. The planar Hall effect is in good agreement with the measured AMR indicating the sample is approximately in a single domain state throughout most of the magnetisation reversal, with a two-step magnetisation jump ascribed to domain wall nucleation and propagation.