MBE 방법에 의한 GaMnAs 및 GaMnN의 성장과 특성 연구 = (A) study on properties and growth of GaMnAs and GaMnN via molecular beam epitaxy

다국어 초록 (Multilingual Abstract)

Ferromagnetic Ⅲ-Ⅴ semiconductors have recently attracted great attention due to their applicability to spin injection devices. While the ferromagnetic metal (FM)-semiconductor junctions show little or no spin injection due to the so-called dead interface layer^(48) the fenomagnetic semiconductor (FS)-semiconductor systems are expected to be absent of such problem. Thus enhanced high injection efficiency in FS -semiconductor junction have recently been demonstrated^(10.49).
GaMnAs is one of the promising Ⅲ-Ⅴ ferromagnetic semiconductors intensively studied by Ohno's group^(50.56)and others^(38,51-52) Ohno's group showed that Mn concentration can be reached >7% by employing low substrate temperatures (T_(s)) cf 200-300℃ in MBE^(5). The main purpose of employing low T_(s), iS to overcome the solid solubility limit of Mn into GaAs lattice at high temperatures.
However, the detail procedure to reach the ultimate goal of growing the high Currie temperature (T_(c)) GAMnAs films by MBE has not been reported. Particularly, it is not clear whether their results were from an optimized condition or not. Namely, it is unclear how they ultimately select the substrate temperature, 2nd what arsenic pressure they used. For the optimization of the growth condition at such low temperatures, one needs to optimize the bask MBE parameters such as As_(4) pressure and Ga flux. As the sticking coefficient of the elements, particularly of As_(4), increases at low T_(s), employing smaller Ⅴ/Ⅲ ratio is anticipated for good crystallographic films.^(53) Although a detail reflection high-energy electron diffraction(RHEED) study in low temperature GaAs growth was reported^(50), it was riot related to optimizing the GaMnAs growth.
Double crystal x-ray diffraction (DCXRD technique has been used for the measurement of Mn concentrations in GaMnAs.^(38,54) Since XRD measurements showed that the GaMnAs films grown in MBE are fully strained,^(54) the estimation of the Mn content using XRD analysis needs appropriate cautions. A simple application of Vegard's law may mislead the estimation. We present in this paper a detail systematic approach to acquire an optimum growth condition for GaMnAs magnetic semiconductors. In-situ RHEED study as well as ex-situ XRD and conductivity measurements were used for this process The coherency among the observations from the different tools is critically described. DCXRD measurements on GaAs films grown at various P_(As) at conventional high temperature showed little difference, and therefore, triple axis HRXRD technique was employed. The HRXRD ω-scan, however, neither differentiates the peaks in the As_(4) pressure range of 1.5∼4.0×10^(-6) torr The full width at half maximum (FWHM) of the (004) peaks was ∼16.6 arcsec, which is very close to that for the substrate, ∼15.5 arcsec. This indicates that GaAs of excellent crystallographic quality can be grown in rather broad range of the As4 pressure. The HRXRD ω-scan on the low-temperature grown GaAs films, however, showed the best FWHM of ∼16.9 arcsec at the condition of T_(s)=250℃ and P_(As)=0.8×l0^(-6) torr. This comparison shows that the structural quality of the grown films little degrades by lowering T_(s), while it becomes more sensitive to the As4 pressure at low T_(s). The results also indicate that the strain parallel to the plane is very small in the films and insensitive to the growth temperature.
More useful information could be obtained from ω-2θ scan measurements on the low-temperature grown GaAs films. The ω-2θ scan enables vertical stress analysis of the films. The Pendellosung fringes reveal high crystalline quality of the grown films. The negative shift of the shoulder peaks indicates compressive stress in the grown GaAs layers. It was shown that the As_(4) pressure for the smallest vertical stress is 8×10^(-7)torr at the given substrate temperature, Thus, it can be concluded that the structure-wise optimum growth can be made at P_(As)=8×10^(-7)torr for the condition of the growth rate 0.25 ㎛/hr and T_(s)=275℃. This rather How optimum arsenic flux relative to that at the high temperature is due partly to an increase of the As_(4) sticking coefficient for stoichiometry at the low temperature, and partly to the reduced growth rate.
For the growth of GaMnAs structures, T_(s) gas varies is the range of 215-275℃ while PAs was fixed at 1.4×10^(-6)torr. The condition for the minimum stress in the GaAs films, or P_(As)=0.8×10^(-6)torr, was found to show spotty RHEED patterns when the Mn cell is open. This suggests that the As_(4) pressure for GaMnAs growt.h needs to be increased to avoid MnAs segregation, Thus employed pressure of 1,4×10^(-6)torr was also expected to enhance Mn incorporation into the Ga sites. In guiding the setting of the Mn cell temperatures, observation of RHEED patterns during growth was very useful. Unless the Mn flux is excess, streaky reconstruction patterns were always observable. We will see this cell temperature achieves the maximum conductivity in the film. Similar streaky patterns were servable till the growth with a little higher (+10) Mn cell temperature, and spotty patterns showed up at further higher Mn cell temperature. In such excess Mn flux conditions, the pattern often changed to a ring pattern after prolonged growth and the films often revealed dark colors with lusterless surface with rapid resistivity increasing. Since high Mn incorporation, while maintaining metallic conductivity, is important for high T_(c) GaMnAs, we tried the Mn flux so far as such segregation patterns did not show up.
The lower conductivity in the below region of a maximum is due to less Mn concentrations, and that in the above region due to compensation and/or MnAs segregation. The higher Mn atoms can be incorporated as the substrate temperature is lower. In other words, the allowed maximum Mn flux, without MnAs segregation, was higher at lower T,. The overall maximum conductivity, however, was observed at T_(s)=250℃. This indicates that not all the incorporated Mn atoms become activated to dopants, particularly at substrate temperatures lower than 250℃. The GaMnAs film of the highest conductivity also was related to the highest ferromagnetic transition temperature. It was revealed that the transition temperature varies linearly with the conductivity at room temperature. Thus, we empirically found that the higher the conductivity of the film is, the higher the ferromagnetic transition temperature is. The resistivity measurement can be a practical tool for forecasting T_(c) of the grown GaMnAs films. Note also that we could correlate the surface reconstruction patterns with the room temperature electrical conductivity of the GaMnAs films.
The best ferromagnetic transport property of the grown GaMnAs semiconductors are T_(s)=250℃, P_(As)=1.4×10^(-6)torr, and T_(Mn)=840℃. The magnetization measured with varying temperature using SQUID revealed the ferromagnetic transition temperature of ∼80 K. We show a clear hysteresis loop in the magnetization-magnetic Geld measurement at 5 K. The Mn concentration calculated by DCXRD was 3.7% for this GaMnAs film.
GaMnN may be one of the possible candidates for room temperature operation,'" and its room temperature ferromagnetism has been reported.^(47,55-56) GaMnN is grown via molecular beam epitaxy (MBE) using a single GaN precursor of (Et_(2)Ga(N_(3))NH_(2)CH_(3)).
The surface morphology of the grown GaMnN layers were flat and smooth with average RMS surface roughness of-4 nm as determined by an atomic force microscopy (AFM) although transmission electron microscopic (TEM) examination revealed a columnar grained hexagonal structure with c-axis alignment. The XRD measurements on GaMnN grown on GaAs(100) revealed clear (0002)peaks indicating a growth of hexagonal c-axis aligned GaN (and GaMnN) structure. Even with such polycrystalline character of the films, doping of Mn could be made successfully in the range of 4-30 a/o as measured by EPMA. Peaks for the 2^(nd) phase appeared at above ∼6 a/o Mn incorporation as determined from other series of experiments. The lattice constant of the 2^(nd) phase Calculated from (111) peaks is 0.389 0,001 nm and fairly well agrees with that of Mn_(3)GaN. 0.390nm^(57) without particular trend with Mn increase. Moreover, the magnetic behavior well conforms to Mn_(3)GaN as will be shown. While it needs further examination, the 2^(nd) phase is tentatively assigned to paramagnetic Mn_(3)GaN at room temperature. Increase of the peak intensity for Mn_(3)GaN (111) is noticeable with diminishing of that for GaN (and GaMnN). Additional peaks for MnsGaN (110) and (220) begin to show up at far higher Mn flux of -30 a/o Mn. This XRD observation coincides well with SQUID measurement results.
From the SQUID results, the magnetization at room temperature caused by GaMnN region of the films increases with Mn_(3)GaN phase segregation. The conductivity also increases with Mn incorporation suggesting again that the room temperature conductivity in magnetic semiconductor can be an indication of the strength of ferromagnetism.^(58) The layers with Mn_(3)GaN phase revealed metallic behavior, but the homogeneous GaMnN layers showed semiconducting or insulating. Therefore, it was difficult to measure magneto-transport property with the homogeneous GaMnN due to the high resistances of the layer and poor electrical contacts. In the meantime, an anomalous Hall effect can be clearly detected from the GaMnN with segregation due to an assistance of the highly conductive Mn_(3)GaN phase. Since the 2^(nd) phase is nonmagnetic at room temperature, the anomalous Hall effect at 300K should come from the GaMnN region of the layer. We speculate that the higher saturation magnetization at room temperature due to GaMnN region in the segregated GaMnN than in the homogeneous GaMnN comes from a high density of carriers originating from Mn_(3)GaN. This result qualitatively support the mechanism of carrier-mediated ferrornagnetism.^(59)
Note also that the conductivity increase with segregation is a very interesting feature of GaMnN in contrast to GaMnAs, where the segregation of metallic MnAs phase caused an resistivity increase of the GaMnAs films.^(58-59) This difference may be explained by a Schottky barrier model, namely metallic MnAs clusters form Schottky barriers depleting carriers from the surrounding GaMnAs region, and thus, increasing the resistance of the film due to reduced current-passing cross-section area of GaMnAs. For GaMnN, however. the segregated metallic Mn_(3)GaN may not form such carrier-depleting energy barrier with respect to the surrounding GaMnN, but rather enhance the overall conductivity by an ohmic contact. While the phenomenological difference has to be further explored, the contribution of the 2^(nd) phase to the conductivity increase in GaMnN layers gave a remarkable impact on investigation of the magnetotransport property in the films. This composition of GaMnN having highly conductive 2^(nd) phase can be practically useful to magnetotransport devices, which require practically good contact fabrication and current pass.
All the approaches to develop Ⅲ-Ⅴ ferromagnetic semiconductors SO far have been focused on increasing the magnetic elements that simultaneously induce magnetic ordering and free carriers. Also, since the room temperature resistivity of a film could be taken as a measure of ferromagnetism and magnetotransport properties anticipated in GaMnN,^(56,58-59) it would be interesting to find out the effects of p-type doping of other than ferromagnetic element in the relation to the ferromagnetism and magnetotransport of the films. This study of co-doping of nonmagnetic element such as Mg with Mn may devote to deeper understanding of the carrier-mediated ferromagnetism in Ⅲ-Ⅴ magnetic semiconductors. GaMnN:Mg layers were grown on sapphire (0001) at a substrate temperature (T_(s)) of 650℃ using MBE. The detail of the growth using this single precursor will be discussed elsewhere, but GaN growth with similar single precursor source was reported by the author.^(9)
From the XRD patterns of Mg co-doped GaMnN layers, we observe (0002) and (0004) peaks for GaN (and GaMnN) in all the samples indicating a c-axis oriented epitaxial growth of the layers on sapphire. The dominant peak is Mn_(3)GaN (200) rather than (111), which is the major peak for the precipitate phase in the films grown on GaAs (100).^(56) One observes that the peak intensity for the Mn_(3)GaN phase decreases with Mg flux. Actually the Mn concentration measured by EPMA has dramatically decreased from ∼3% for GaN:Mn layer down to ∼0.3% in GaN:(Mn,Mg) layer for T_(Mg)= 405℃ although the Mn impingement rates were controlled the same. It is also interesting to note that segregation of Mn_(3)GaN phase occurred at such low Mn concentrations, where GaN:Mn layers revealed homogeneity if Mg is not co-doped. These results indicate that Mn and Mg competes each other to find Ga sites at impingement on the substrate surface during the growth. Furthermore, Mg incorporation promotes local aggregation of Mn to form Mn_(3)GaN clusters.
The Mg concentration increased linearly with Mg flux under the given Mn flux. While the conductivity has dramatically increased, the layers still revealed semiconducting behaviors. This trend was the same for Mg-codoped homogeneous GaMnN layers. It is very interesting to compare the results with the resistivity of GaN doped purely with Mn. This comparison indicates that, in the investigated range of doping. Mg behaves of the simple ionizing impurities in semiconductors, but Mn contributes to metal-insulator transition as well although the ability to offer conduction carriers is less. Although fen~magnetism was confirmed for both cases, anomalous Hall effect was confirmed only for the GaN:Mn layer^(56) while it was not clearly identified from the GaN;(Mn,Mg) layers are room temperature. In the discussion, note that the conductivity is the average value of the films containing the GaMnN or Ga(Mn,Mg)N solid solution region and the precipitates, but the room temperature ferromagnetism comes only from the GaMnN or Ga(Mn,Mg)N solid solution region of the films. The observed higher saturation magnetization (M_(s)) for the GaN:Mn layer, even with lower carrier concentration, suggests the importance of the critical level of Mn concentration in the carrier-mediated ferromagnetism In GaMnN.
For the homogeneous Ga(Mn,Mg)N layers with less Mn, the effect of increasing carrier concentration is small suggesting again the importance of the amount of Mn to enhance the magnetic property. Note, however, also that the M_(s) has increased with Mg. This is remarkable because the M_(s) increase was realize with reducing Mn content along with Mg increase. However, one can see that the Mn atoms participating in ferromagnetic coupling at room temperature are greater (or those paaricipating in antiferromagnetic coupling are less) in the layers with smaller Mn concentration but with higher overall hole concentration. It is a remarkable observation that Mg codoping enhances the participation of Mn into ferromagnetic coupinh, but the origin need to be explained. In this case, the participation of Mn is-μB. Nevertheless, this clearly indicates that the holes from Mg interact with Mn to enhance carrier-mediated ferromagnetism in GaMnN.
From these results, we successfully grew GaMnAs and GaMnN thin films on GaAs(100) and sapphire substrates at low temperatures via molecular beam epitaxy. We observe ferromagnetism and magnetotransport for GaMnAs at low temperature(<80K), and for GaMnN at room temperature. We can understand the carrier-mediated ferromagnetism in Ⅲ-Ⅴ magnetic semiconductors from co-doping of nonmagnetic element such as Mg with Mn.

목차 (Table of Contents)

• 목차 = Ⅰ
• 제1장 서론 = 1
• 제2장 이론적 배경 = 6
• 2.1. 자성반도체 = 6
• 2.1.1. Diluted magnetic semiconductor = 6
• 2.2. 자성반도체의 강자성 특성 이해 = 11
• 2.1.1. Mn 이온의 전하량과 스핀상태 = 11
• 2.2.2. 금속과 부도체 사이의 상전이 근처 전자 상태 = 12
• 2.2.3. Carrier에 의해 매개되는 강자성 모델(Zener model) = 14
• 2.3. 자성반도체의 전기적 특성 이해 = 15
• 2.3.1. 온도의존성 비저항 = 15
• 2.3.1.1. 금속에서의 온도의존성 비저항 = 15
• 2.3.1.2. Oxide에서의 온도의존성 비저항 = 16
• 2.3.1.3. 묽은 자성 불순물에 의한 비저항 = 16
• 2.2.2. 전기 비저항 = 17
• 2.3.2.1. 천이 금속에 대한 two-current 모델 = 24
• 2.3.2.2. 불순물 영향 = 25
• 2.3.2.3. 온도 의존성 = 26
• 2.3.3. Galvanomagnetic Effects = 31
• 2.3.3.1. 비자성 재료:정상 Hall 효과와 자기 저항 = 31
• 2.3.3.1.1. 정상 Hall 효과 (Ordinary Effect) = 31
• 2.3.3.1.2. 자기저항(magnetoresistance) = 32
• 2.3.3.2. 강자성 재료에서 Galvanomagnetic effects = 32
• 2.3.3.2.1. 비정상 Hall 효과 = 33
• 2.3.3.2.2. Planner Hall 효과 = 35
• 제3장 GaMnAs 성장과 특성 = 41
• 3.1. 서론 = 41
• 3.2. 실험방법 = 41
• 3.2.1. MBE 시스템 = 41
• 3.2.2. 기관 세척 = 42
• 3.2.3. Source flux calibration = 42
• 3.2.4. 기판 온도 calibration = 49
• 3.2.5 성장 순서 = 50
• 3.3. 결과 및 고찰 = 54
• 3.3.1. 고온 GaAs 에피층 성장 = 54
• 3.3.2. 저온 GaAs 에피층 성장 = 55
• 3.3.2.1. As 압력 변화에 의한 GaAs 성장 = 55
• 3.3.3. 저온 GaMnAs 성장 = 61
• 3.3.3.1. 저온 GAMnAs 성장 = 61
• 3.3.3.2. GaMnAs의 구조적 특성 = 65
• 3.3.3.3. GaMnAs의 전기적 특성 = 75
• 3.3.3.3.1. As 압력에 의한 전도도 = 75
• 3.3.3.3.2. 기관 온도 변화에 의한 전도도 = 75
• 3.3.3.3.3. Mn flux 변화에 의한 전도도 = 75
• 3.3.3.3.4. 전기 전도도와 자성 특성 = 77
• 3.3.3.4. GaMnAs의 자기적 특성 = 85
• 3.3.3.4.1. Mn 함량에 따른 M-H 특성 = 85
• 3.3.3.4.2. Mn 함량에 따른 M-T 특성 = 86
• 3.3.3.4.3. GaMnAs의 비정상 Hall 효과 = 86
• 3.3.3.4.4. GaMnAs의 planar Hall 효과 = 89
• 3.4. 결론 = 101
• 제4장 GaMnAs 성장과 특성 = 102
• 4.1. 서론 = 102
• 4.2. 실험 방법 = 102
• 4.2.1. GaN 단일 전구체 및 성장 순서 = 102
• 4.2.1.1. 선정 배경 및 재료 = 102
• 4.2.1.2. 실험 순서 = 103
• 4.3. 결과 및 고찰 = 104
• 4.3.1. GaN 단일전구체를 이용한 GaN 성장 = 104
• 4.3.1.1. 기관 온도 변화에 따른 XRD 결과 = 104
• 4.3.1.2. 성장 압력 변화에 따른 XRD 결과 = 105
• 4.3.1.3. 성장 온도 변화에 따른 AFM 결과 = 109
• 4.3.2. GaMnN의 성장과 특성 = 112
• 4.3.2.1. GaMnN 성장 = 112
• 4.3.2.2. GaMnN의 구조적 특성 = 113
• 4.3.2.2.1. Mn flux 변화에 의한 GaMnN의 XRD 결과 = 113
• 4.3.2.2.2. 기관 온도 변화에 의한 GaMnN의 XRD 결과 = 121
• 4.3.2.3. GaMnN의 전기적 특성 = 128
• 4.3.2.3.1. Mn 함량에 따른 전기 전도도 = 128
• 4.3.2.4. GaMnN의 자기적 특성 = 133
• 4.3.2.4.1. Mn 함량에 따른 M-H 특성 = 133
• 4.3.2.4.2. Mn 함량에 따른 M-T 특성 = 134
• 4.3.2.4.3. GaMnN의 비정상 Hall 효과 = 135
• 4.3.2.4.4. GaMnN의 planar Hall 효과 = 137
• 4.3.3. GaMnN:Mg 성장 및 특성 = 146
• 4.3.3.1. GaMnN:Mg 박막의 성장 = 146
• 4.3.3.2. GaMnN:Mg의 구조적 특성 = 146
• 4.3.3.3.2.1. Mg flux 변화에 의한 XRD 결과 = 146
• 4.3.3.3. GaMnN:Mg의 전기적 특성 = 151
• 4.3.3.3.1. Mg flux 변화에 의한 전도도 = 151
• 4.3.3.3.2. GaMnN:Mg의 온도의존성 비저항 = 151
• 4.3.3.4. GaMnN:Mg의 자기적 특성 = 156
• 4.3.3.4.1. Mg flux 변화에 의한 M-H 특성 = 156
• 4.3.3.4.2. Mg flux 변화에 의한 M-T 특성 = 158
• 4.4. 결론 = 163
• 제5장 결론 = 165
• Reference = 167