In situ annealing of III_1-xMn_xV ferromagnetic semiconductors Free

Considerable effort has been devoted to the study of III_1-xMn_xV magnetic semiconductors due to their potential for applications in spintronic devices.^1,2 It is already well established that the ferromagnetic interaction between magnetic ions in these materials is mediated by holes, and that point defects such as Mn interstitials (Mn_I) play a crucial role in determining the magnetic properties of the III_1-xMn_xV alloys.³ Earlier experimental studies have shown that the Curie temperature of Ga_1-xMn_xAs epilayers can be further increased by postgrowth heat treatment, i.e., by low-temperature (LT) annealing.^4,5 Specifically, channeling Rutherford backscattering and channeling particle-induced x-ray emission experiments revealed that LT-annealing leads to out-diffusion of the highly mobile Mn_I defects from the specimen bulk, accumulating at the free surface of the film.⁶ Such removal of Mn_I from the bulk then leads to increases of the Curie temperature T_C, of the carrier concentration p, and of the average magnetic moment of the Mn content through.^6,7 Annealing experiments on Ga_1-xMn_xAs epilayers were typically carried out ex situ, in air or in nitrogen gas, or in situ under As or Sb capping, at different annealing temperatures and for different annealing times.^4,5,8–10 This approach typically involves surfaces terminated with an oxide or with MnAs(Sb) clusters, which automatically inhibits further overgrowth on top of the annealed Ga_1-xMn_xAs film.^11,12

In this work, we carried out a systematic study of annealing III_1-xMn_xV films (Ga_1-xMn_xAs and Ga_1-xMn_xAs_1-yP_y) in situ, i.e., without removal from the molecular beam epitaxy (MBE) chamber, with different capping layers (e.g., As, Se, or Te). Our results show that a correct in situ LT-annealing approach can lead to increases of T_C, p, and the magnetic moment, as in the earlier optimized ex situ annealing experiments. Importantly, however, this approach also allows deposition of high-quality semiconductor layers directly on top of such in situ annealed films, demonstrating great potential for designing high quality III_1-xMn_xV-based multilayers for spintronic applications, that take advantage of the benefits of the annealing process.

In order to systematically investigate the effects of in situ annealing on structural and magnetic properties in III_1-xMn_xV films, we grew by MBE two series of samples on GaAs (100) substrates: 12 nm thick Ga_1-xMn_xAs films and 10–50 nm thick Ga_1-xMn_xAs_1-yP_y films,¹³ with Mn concentrations ranging between 4 and 8%, and P concentrations of about 15%. High-purity Ga and Mn elemental fluxes were provided by standard effusion cells, and As₂ and P₂ fluxes were generated by cracker cells. The growth was monitored by in situ reflection high energy electron diffraction (RHEED), which shows that the RHEED of GaMnAs or GaMnAsP epilayers is of the (1 × 2) type, and LT-AlGaAs shows (1 × 1) pattern (no reconstruction), indicating that the grown films are monocrystalline.¹⁴ Epi-ready GaAs semi-insulating (100) substrates from AXT, Inc., are used in this study. The background residual pressure in the MBE chamber is below 1 × 10⁻⁹Torr. The substrate temperature is measured by a thermocouple which was carefully calibrated at the melting points of indium and at the temperature of oxide desorption from the GaAs surface. After being placed in the preparation chamber, the GaAs substrates were cleaned in situ by oxide desorption, by heating to 600 °C. The completion of the cleaning process was recognized by monitoring the RHEED pattern. The procedure for MBE growth of the GaMnAs or GaMnAsP was as follows. First a GaAs buffer of thickness 100–200 nm was grown at high temperature, T_S ∼ 600 °C (i.e., under normal GaAs growth conditions). Then the substrate was immediately cooled down to a temperature of 250 °C for LT growth. Using As₂:Ga beam equivalent pressure ratio of 20:1, the GaMnAs or GaMnAsP layers were then grown at this lower temperature. After growth, the films were cooled to room temperature, and the in situ annealing procedure was carried out as follows. An amorphous capping layer of As, Se, or Te was first deposited in MBE vacuum chamber to thickness of a few tenths of a nanometer. The substrate temperature of the film was then gradually increased to 200 °C and then to 300 °C in a two-step annealing process, with 15 min per step, without removal from the MBE chamber. During annealing, RHEED showed the recovery of a (1 × 2) streaky RHEED pattern of the III_1-xMn_xV films after desorption of amorphous capping layer. For comparison, a sample without a capping layer was also annealed by the same procedure, which we will refer to as “vacuum-annealed.” Additionally, we grew a III_1-xMn_xV sample with a AlGaAs capping layer on such in situ annealed III_1-xMn_xV film. Such overgrowth by high-quality semiconductor layers on top of in situ annealed films demonstrates great potential for designing high quality III_1-xMn_xV-based multilayers for spintronic applications, which take advantage of the benefits of LT annealing of the constituent ferromagnetic layers.

Magnetization measurements on the III_1-xMn_xV films obtained by this process were carried out as a function of magnetic field and temperature using a Quantum Design superconducting quantum interference device (SQUID) magnetometer. X-ray diffraction (ω-2θ) measurements were also performed to calculate the corresponding out-of-plane lattice constants. Resistivity was measured as a function of temperature in order to examine the effects of annealing on the hole concentration. Finally, x-ray photoelectron spectroscopy (XPS) measurements were carried out on the in situ Se-annealed GaMnAs/AlGaAs heterostructure. After the spectrum was measured, the surface was bombarded in situ by Ar⁺ ions accelerated to 1.4 keV for 10 min at 5 × 10⁻⁶ mbar. The relevant high resolution core line spectra of the bombarded surface, e.g., Al 2p level, Mn 2p_3/2 level, and Se 3d level, were repeatedly collected seven times to probe the compositional profile of the specimen.

In Table I, we list the structural parameters of the GaMnAs films obtained under various annealing processes. Note that sample A* and E* are parts from samples A and E, respectively. All samples were investigated by SQUID magnetometry. The magnetic parameters, such as the Curie temperature T_C and saturation magnetization M_s are also obtained by SQUID measurements, are also listed in Table I. Hysteresis loops at T = 5 K along the [100] crystallographic direction for samples A, A*, B, and C are shown in the top panels of Fig. 1. The hysteresis loops of samples D and E (not shown) are similar to that of sample A except for their magnitude; the hysteresis loop for sample E* (not shown) is similar to that of sample A*. The corresponding temperature dependences of magnetization measured in a 10 Oe field applied along the [100] direction is shown in the bottom panels of Fig. 1. Note that all annealed samples have a higher Curie temperature and a higher magnetization, clearly indicating that the overall quality of the ferromagnetic order is enhanced by the in situ LT-annealing process. Note also that the LT-annealing in vacuum without the amorphous capping layer does not have such dramatic effect, indicating that the amorphous Se capping layer plays an important role in the annealing process. Moreover, it should be pointed out that we have achieved a high Curie temperature (T_C ∼ 133 K) in a 12 nm GaMnAs/12 nm AlGaAs heterostructure (sample B). Interestingly, hysteresis loop of the latter sample appears to be modified by AlGaAs overgrowth compared with that of sample C, which remains to be investigated. To further investigate the electrical properties of in situ-annealed GaMnAs samples, electrical transport measurements were performed on both as-grown and annealed specimens. During the measurements, the specimen was loaded into a Janis closed-cycle helium flow cryostat with a temperature range between 12 and 270 K. Temperature-dependent sheet resistivity for GaMnAs samples (B, C, D, and E) are shown in Fig. 2. The resistivity data for sample A (not shown) are similar to those of sample D.

All of the samples show a clear resistivity maximum point T_ρ, which can be regarded as an estimate of T_C (usually slightly higher than T_C).¹⁵ Curie temperatures obtained from the ρ-T curves are in good agreement with those measured by SQUID. In addition to the enhancement of T_C, in situ Se-annealing leads to lower sheet resistivities because of the removal of Mn_I from the sample bulk. Note that after Se-annealing, ρ at low temperature becomes smaller than ρ at room temperature, indicating that the material has undergone a transition from a semiconductorlike type to a metallic type.

The above results indicate that in situ LT-annealing (i.e., without removal from the MBE chamber) with amorphous Se capping layers leads to an increase of T_C, of magnetic moment, and of p (the latter resulting in a semiconductor-to-metallic transition), similar to findings in earlier optimized ex situ annealing experiments. It is widely accepted that such LT-annealing leads to the removal of Mn atoms from interstitial positions Mn_I. It is also well known that the interstitial Mn_I ions are positively charged double donors and are thus attracted to the negatively charged substitutional Mn acceptors Mn_Ga to form Mn_Ga–Mn_I pairs. Theoretical calculations¹⁶ showed that the Mn_I atoms do not contribute to the ferromagnetic coupling mediated by free holes and, furthermore, that the Mn_Ga–Mn_I pairs are coupled antiferromagnetically through superexchange interaction, so that they do not contribute magnetization. As a result, the increase of the hole concentration p (measured by the electric transport), as well as the increase in saturation magnetization and the increase of the Curie temperature observed on annealed samples all serve to corroborate that in situ Se-annealing results in the removal of a significant fraction of Mn interstitials in GaMnAs.¹⁷ 

Figure 3 shows changes in the XPS spectra arising from the Ar-ion etching of the surface of in situ Se-annealed GaMnAs/AlGaAs heterostructure. Noticeable intensity changes of the core lines were observed in high-resolution XPS spectra for a sequence of etchings. The arrow in the figure indicates the interface between GaMnAs and AlGaAs. Figure 3 shows the Al 2p, Mn 2p_3/2 and Se 3d core level spectra. Al 2p peak centered at 73 eV, whose intensity decreases dramatically after 40 min of Ar-ion etching, corresponding to the appearance of Mn 2p_3/2 peak centered at 639 eV. As shown in Fig. 3, a Se 3d core level peak located at 53.8 eV appears only at the interface between GaMnAs and AlGaAs. This indicates the presence of a Se-related compound at the interface, although we have observed the desorption of Se during the LT-annealing process. We speculate that the removal of Mn_I is assisted by a Se related solid–solid reaction occurring on the surface of GaMnAs, similar to what occurs during ex situ LT-annealing, where the GaMnAs surface is passivated with an oxide.

For completeness, we have fabricated a series of GaMnAsP under in situ annealing process with various amorphous capping layer. Table II listed the structural of GaMnAs films under various annealing process. Note that, sample G* is part piece from sample G. All samples are investigated by SQUID magnetometry and XRD. The obtained magnetic parameters such as Curie temperature T_C and saturation magnetization M_s are also listed in the table.

Hysteresis loops at T = 5 K along the [001] axis for samples G, I, J, and K are shown in the top panels of Fig. 4. The corresponding temperature dependence of magnetization measured in a 10 Oe field applied along the [001] direction is shown in the bottoms panels of Fig. 4. Note that all annealed samples has the higher Curie temperature and much square hysteresis loops, indicating that the overall quality of ferromagnetic order is enhanced by LT-annealing process with As, Se, and Te amorphous capping layer, indicating that the amorphous capping layer plays an important role in annealing process. The 2θ-ω coupled XRD scans (not shown) were also carried out on every sample. Without significant changes to the intensity of peaks, it is noted that the diffraction spectrum from the LT-annealed epilayers is shifted to the right by about 0.21° compared with that measured for the as-grown sample, from 66.36° to 66.57° for 2θ at the (004) reflection. This indicates that the value of the out-of-plane lattice constant is reduced from 5.630 to 5.614 Å, thus suggesting the elimination of interstitial Mn atoms in GaMnAsP samples by the annealing process.

The in situ annealing effect on magnetic properties of ternary alloy Ga_1-xMn_xAs and quaternary alloy Ga_1-xMn_xAs_1-yP_y on GaAs (100) substrate is explored. We discovered that the in situ annealing has the similar effect on III-Mn-V as the ex situ annealing, provided that an amorphous cap layer of As, Se, or Te is deposited at room temperature first before the low-temperature annealing. XPS spectra analysis suggests that the effect of low temperature annealing on III_1-xMn_xV might be due to the diffusion out of “high mobility” Mn interstitials Mn_I which is driven by the solid–solid reaction between Mn ions and the elements (As, Se, or Te; O in ex situ case) on surface below 300 °C. Importantly, in situ Se annealing enables us to fabricate the heterostructures (e.g., modulation doping heterostructure) with the embedded high quality GaMnAs layers.

This work was supported by the NSF Grant No. DMR14-00432, Basic Science Research Program through the NRF of Korea funded by the Ministry of Education (No. 2015R1D1A1A01056614), and a grant from Korea University. P.S. and S.P. acknowledge the U.S. Department of Energy Office of Science, Office of Basic Energy Sciences, under the Award No. DE-FC02-04ER15533 (No. NDRL No. 5200).