Spin Transport in Ge Nanowires for Diluted …drl.ee.ucla.edu/wp-content/uploads/2017/07/spin...Spin Transport in Ge Nanowires for Diluted Magnetic Semiconductor-based Nonvolatile

Spin Transport in Ge Nanowires for Diluted Magnetic Semiconductor-based Nonvolatile Transpinor

Jianshi Tang, Tianxiao Nie, and Kang L. Wang

Device Research Laboratory, Department of Electrical Engineering, University of

California, Los Angeles, California 90095, USA

Spintronic devices, in particular spin field-effect transistors (spinFETs), have been researched for decades as a promising candidate to replace Si transistors with potentially low power dissipation and low variability. In this paper, we propose a nonvolatile spin-base transistor (transpinor) based on diluted magnetic semiconductor (DMS) with electric field-controlled paramagnetism-to-ferromagnetism phase transition. To realize the transpinor, we demonstrate the electrical spin injection into Ge nanowires using both ferromagnetic Mn5Ge3 Schottky contacts and Fe/MgO tunnel junctions. We observe much longer spin lifetimes and diffusion lengths in Ge nanowires compared with bulk Ge. Furthermore, we successfully grow the single-crystalline MnxGe1-x DMS nanowires using pattern-assisted molecular beam epitaxy (MBE). The MnxGe1-x nanowires exhibit a Curie temperature above 400 K, and the ferromagnetism can be further modulated by an external gate voltage. The demonstrated electric field-controlled ferromagnetism, along with the successful spin injection into Ge nanowires, paves the road to build the proposed transpinor.

Introduction

As the major economic growth driver for decades, the continuous scaling of the feature size in the Si technology is approaching the ultimate physical limit of several nanometers. Several critical challenges are highlighted in the International Technology Roadmap of Semiconductors (1), including the increase in functionality, and the decrease in power dissipation and manufacturing variability. Novel materials and devices are in urgent need to resolve those critical challenges. Spintronics (or spin-based electronics) have emerged as a promising solution by utilizing the spin of electrons as another degree of freedom for information processing, which could enable the creation of novel electronic devices with increased functionalities, low power dissipation, and nonvolatility (2-3). In particular, several prototypes of spinFETs have been proposed and extensively studied as an appealing substitute for Si complementary metal-oxide-semiconductor (CMOS) devices (4-6). The operation of spinFETs typically involves the injection, manipulation and detection of electron spins in the spin/charge transport process. For example, the Datta-Das-type spinFET uses a gate electrode to modulate the spin state of the injected carriers through the Rashba effect (4). Another variant, proposed by Sugahara (5), is a spin-MOSFET comprised of a MOS capacitor with half-metallic source/drain contacts, in which the gate electrode modulates the spin-dependent barrier at the source/drain junctions. More recently, Behin-Aein et al proposed an interesting all-

10.1149/06406.0613ecst ©The Electrochemical SocietyECS Transactions, 64 (6) 613-623 (2014)

613) unless CC License in place (see abstract). ecsdl.org/site/terms_use address. Redistribution subject to ECS terms of use (see 128.97.89.222Downloaded on 2015-02-17 to IP

http://ecsdl.org/site/terms_use

spin logic device with built-in memory that uses spin at every stage of its operation to avoid repeated spin-to-charge conversion, which could help reduce the power dissipation (6). Nevertheless, all of the above approaches still use electronic charge current that inevitably consumes power. As a result, the potential advantage in reducing the power dissipation is very limited compared with conventional CMOS devices. Propose of DMS Nanowire-based Nonvolatile Transpinor

In order to make a spinFET that can truly compete with Si CMOS device, one

practical approach is to reduce or eliminate the charge current flow. Here we propose a novel transpinor based on DMS nanostructures whose ferromagnetism can be controlled by an external electric field. The use of DMS may provide a promising solution to minimize the charge current flow. It allows us to manipulate the magnetic state of the DMS channel through a voltage signal without current flow in the channel (see Figure 1). In this transpinor device, the magnetic moments are transferred from the source to the channel, and then to the drain through spin injection and exchange interaction, respectively. As the gate electrode manipulates the ferromagnetism of the DMS channel, it controls the communication between source and drain.

Recently, we have demonstrated that the growth of singe-crystalline Mn-doped Ge DMS quantum dots with a Curie temperature TC above 400 K, and more significantly the electric-field-controlled ferromagnetism up to room temperature (7). In that work, Mn-doped Ge nanostructures were used to enhance the carrier-mediated ferromagnetism and meanwhile minimize material defects. This hence points out a unique advantage for nanostructures over bulk materials in fabricating practical spintronic devices. For the convenience of device fabrication, here we choose Mn-doped Ge DMS nanowire as the channel material, in which the electric-field-controlled ferromagnetism with a high TC is expected to be similar to that in quantum dots. We further integrate the DMS nanowire with high-quality ferromagnetic contacts for spin injection to build the prototype of transpinor. Two possible implementations are schematically illustrated in Figures 1(a-b) with Mn5Ge3 and Fe/MgO as source/drain contacts, respectively.

Unlike the Datta-Das type spinFET relying on the Rashba effect to control the spin precession of individual electron, the proposed transpinor rather manipulates a collection of spins via the carrier-mediated paramagnetic-to-ferromagnetic transition as a single identity and thus is more energy-efficient and robust (8). The use of N collective spins can be treated as a single identity and will enable the ultimate power dissipation of a single switching element (nanomagnet) of kBTln2, instead of NkBTln2 per switch for conventional equilibrium computing (9). Moreover, the input and output information can be stored in the ferromagnetic contacts (nanomagnets); therefore, it inherently provides nonvolatility in our device and hence eliminates the standby power dissipation.

It should be pointed out that one of the key challenges to integrate spintronic devices into CMOS circuits is to achieve efficient conversion between a spin signal (magnetization) and a charge signal (current/voltage), and the concatenability (device fan out). Similarly, in order to cascade our transpinor devices for logic computation, several strategies can be adopted to transfer information from one transpinor device to another, as shown in Figure 1(c). Possible solutions may involve magnetic tunnel junctions (MTJs), spin valves, or the spin Hall effect (10). For instance, CoFe/MgO/Fe(Mn5Ge3) MTJ or CoFe/Cu/Fe(Mn5Ge3) metallic spin valve can be integrated on top of the nanomagnets to read out and also manipulate their magnetizations. Alternatively, the recently discovered

ECS Transactions, 64 (6) 613-623 (2014)



voltage-controlled magnetic anisotropy (VCMA) effect can be used to read/write the magnetization information in the nanomagnets (11). In this way, the control and clocking schemes will be carefully designed to enable the switching of the next stage. The read out voltage signal can be used to drive the switching of the next transpinor stage, which ideally requires a very small range of gate voltage in the vicinity of paramagnetic-to-ferromagnetic phase transition.

Substrate

Metal Gate

Mn5Ge3 Mn5Ge3MnxGe1-x NW

Al2O3

Mn Mn

Drain Source

Metal Gate

Substrate

MnxGe1-x Nanowire

Al2O3

Drain Source

MgOFeFe

(a) (b)

Substrate

Gate

MnxGe1-x DMS

FeFe

Al2O3

MgO

FeFe

Stage 2

Gate

Transpinor

CoFe

Stage 1

Vread

FeFe

GateAl

(2) Inversed VCMA

FeFe

Gate

Vread

Vdd

Cu/MgO

(1) Spin Valve / MTJ

CoFe

(c)

Figure 1. DMS Nanowire-based Nonvolatile Transpinor. (a-b) Proposed diluted magnetic Ge nanowire-based nonvolatile transpinor with Fe/MgO tunnel junctions and ferromagnetic Mn5Ge3 Schottky junctions, respectively. The magnetic moments are transferred from the source to the channel, and then to the drain with a spin gain. The MnxGe1-x DMS nanowire channel, whose paramagnetism-to-ferromagnetism phase transition is electrically controlled by the gate voltage, communicates with the input and output nanomagnets through spin injection/extraction and exchange interaction, respectively. (c) Two possible schemes to convert the magnetization to a voltage signal in the read out of a transpinor circuit. The devices can be cascaded, very much like CMOS, but in the magnetic state. The first scheme adopts either a metallic CoFe/Cu/Fe(Mn5Ge3) spin valve or a CoFe/MgO/Fe(Mn5Ge3) MTJ connected to the power supply Vdd through a resistor. The other scheme uses the voltage-controlled magnetic anisotropy (VCMA) effect, in which the switching in the output magnet will induce a voltage pulse in the read out.




Electrical Spin Injection and Transport in Ge Nanowires In order to realize the proposed transpinor, we first studied the spin transport and

demonstrated the electrical spin injection and detection in Ge nanowires. While the electrical spin injection into ordinary metals can be easily demonstrated in metallic spin valve structures (12-13), it should be pointed out that the realization of efficient spin injection into semiconductors is much complicated by several factors (14): 1) the large difference in the conductivity between ordinary ferromagnetic metals (FM) and semiconductors (SC) would make the spin injection efficiency negligibly small, well known as the conductivity mismatch problem (15); 2) the increase in the doping concentration in order to reduce such a conductivity difference, however would decrease the carrier spin lifetime due to the aggravated spin relaxation from impurity scatterings (2); 3) while the insertion of a tunneling or Schottky barrier helps alleviate the conductivity mismatch (16-17), the preparation of a high-quality tunneling oxide without pinholes or a defects-free Schottky contact without Fermi-level pinning is not trivial (18-19). Moreover, the localized states at the FM/SC interface and the surface roughness could significantly complicate and jeopardize the spin injection process (20). It is also worth noting that the interface states could also induce Fermi level pinning in metal/semiconductor contacts. For instance, the Fermi level in conventional metal/Ge contacts is strongly pinned close to the Ge valence band edge due to a high density of interface states (21). This problem is even worse for Ge nanowires which have a higher surface-to-volume ratio. Therefore, contact engineering is the key to realize high-efficiency spin injection into Ge nanowires.

In our earlier work, we have established a convenient method to fabricate high-quality germanide contacts in Ge nanowire transistors through rapid thermal annealing (RTA), which promoted the solid-state reaction between the Ge nanowire and metal contacts. The salient feature of this process is that the formed germanides are single crystalline with atomically clean interfaces, which help alleviate the Fermi level pinning and also allow for controlled scaling of the Ge nanowire channel length down to tens of nanometers. Therefore, this facile approach offers great advantages over modern high-cost and complex photolithography technology in fabricating short-channel transistors and scaled nanodevices. Using this method, we have demonstrated various single-crystalline germanide contacts, including Ni2Ge, NiGe, Ni3Ge, and Mn5Ge3, for building high-performance Ge nanowire transistors (22-25). More importantly, the fabricated single-crystalline Mn5Ge3 contact with room-temperature ferromagnetism (TC ~ 300 K) and high-quality interface enabled the electrical spin injection into the p-type Ge nanowire through a Schottky barrier (14, 26), as shown in Figure 2. When applying a large bias on the Mn5Ge3/Ge/Mn5Ge3 nanowire transistor, spin-polarized carriers were injected into the Ge nanowire from one ferromagnetic Mn5Ge3 contact (namely the spin injector) and then were scattered as they travelled along the Ge nanowire channel before reaching the other ferromagnetic Mn5Ge3 contact (namely the spin detector). This process is schematically illustrated in Figure 2(b), and yields to a negative and hysteretic magnetoresistance (MR) as shown in Figure 2(e). It was found that the MR of the Mn5Ge3/Ge/Mn5Ge3 nanowire transistor exponentially decayed with the increasing Ge nanowire channel length, as shown in Figure 2(f). Using the modified Julliere’s model (27), a spin diffusion length of λsf = 480 nm and a spin lifetime of τsf = 244 ps were revealed in p-type Ge nanowires at T = 10 K, which were much larger than those reported for bulk p-type Ge with a similar doping concentration (28).




Mn5Ge3MnGe

SiO2

Si

(a)

-30 -20 -10 0 10 20 30 40

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

MR

(%

)

Magnetic Field (kOe)

Lch=700 nm

Lch=550 nm

Lch=450 nm

T = 10 KIdc = 10 µµµµA

450 500 550 600 650 700-1.3

-1.2

-1.1

-1.0

-0.9

-0.8

-0.7

-0.6

ln [

MR

(H//=

30 k

Oe)

]

Channel Length (nm)

T=10 K

λλλλsf = 480±±±±13 nm

(e) (f)

Mn5Ge3

Mn5Ge3

p-GeEg~0.66 eV

EFM

φφφφSB~0.25 eV

eVDS

(b)

(002)

(002)

(111)

(002)

[110]

Ge

Mn5Ge3(200)

[010](002)

- -

Mn5Ge3

Ge10 nm

Ge nanowire

Mn5Ge3nanowire

[111]

(c) (d)

Figure 2. (a) Schematic diagram of the Mn5Ge3/Ge/Mn5Ge3 nanowire transistor. (b) Energy band diagram of the Mn5Ge3/Ge/Mn5Ge3 nanowire transistor under a large dc bias. The transport process of spin-polarized carriers (holes) in the Mn5Ge3/Ge/Mn5Ge3 nanowire transistor is also illustrated. (c) Transmission electron microscopy (TEM) image of the formed Mn5Ge3/Ge nanowire heterostructure upon RTA at 450 oC. (d) High-resolution TEM image of the atomically clean Mn5Ge3/Ge interface. The insets show the diffraction patterns of Mn5Ge3 and Ge, respectively. (e) MR curves of three Mn5Ge3/Ge/Mn5Ge3 nanowire transistor with three different channel lengths (Lch = 450, 550, and 700 nm) at T = 10 K under a dc current bias of Idc = 10 µA. The black and red arrows indicate the backward and forward sweeping directions of the axial magnetic field, respectively. All the MR curves are intentionally offset by multiples of 0.1% for clarity. (f) Semi-log plot of the MR magnitude at H// = 30 kOe versus the channel length for the three Mn5Ge3/Ge/Mn5Ge3 nanowire transistors. The linear fitting (red curve) yields a spin diffusion length of λsf = 480±13 nm in the p-type Ge nanowire at T = 10 K.




Meanwhile, tunneling spin injection into n-type Ge nanowires was demonstrated using epitaxial Fe/MgO contacts, which were grown by molecular beam epitaxy (MBE), as shown in Figure 3. It should be pointed out that the cylindrical geometry of nanowires presents a major challenge in growing defects-free tunnel oxides because of there is no well-defined lattice plane on the round nanowire surface. In an earlier work, we have demonstrated the MBE epitaxial growth of high-quality Fe/MgO tunnel junctions on a Ge wafer (18), and later realized electrical spin injection into bulk Ge (29). The unique 45 degree rotation between the MgO and Ge unit cells helped enhance the symmetry-induced spin filtering and also minimize lattice defects (30). Here we adopted the same technique in fabricating Ge nanowire devices with epitaxial Fe/MgO junctions for tunneling spin injection. We also intentionally increased the MgO thickness (2-3 nm compared with 0.5-1 nm in bulk Ge devices) in order to improve the MgO coverage on the Ge nanowire surface, as shown in Figure 3(b). Temperature-dependent I-V measurements were further performed on the Fe/MgO/Ge nanowire tunnel junctions to investigate the electrical properties, as shown in Figure 3(c). The nonlinear I-V characteristics, along with the weak temperature dependence of the resistance-area (RA) product, affirmed the tunneling nature of the Fe/MgO junction.

To characterize the electrical spin injection into the Ge nanowire through the Fe/MgO tunnel junction, nonlocal spin valve measurements were performed, as shown in Figure 3(d). In this measurement, spin-polarized carriers were injected from one ferromagnetic contact (spin injector) to create a spin accumulation in the Ge nanowire channel. As the spin-polarized carriers diffused away from the spin injector, another ferromagnetic contact (spin detector) was placed outside the charge current loop to probe the spin-dependent electrochemical potential of a particular spin channel (spin up or spin down, depending on the magnetization of the spin detector) related to the reference contact. An in-plane magnetic field was swept along the easy axis of the ferromagnetic spin injector and detector to change the relative magnetization direction of the two. As the relative magnetization direction of the spin injector and the detector changed between parallel and anti-parallel states, a bipolar nonlocal voltage VNL was sensed.

Figure 3(d) shows a typical nonlocal spin valve signal in n-type Ge nanowire at T = 40 K, in which the relative magnetization direction of the spin injector and detector was labeled under different magnetic fields. The spin diffusion length of λsf = 2.57 µm and the spin lifetime of τsf = 7.2 ns were obtained, which were again much larger than those in bulk n-type Ge with a similar doping concentration (29). The significant enhancement in the spin lifetime and diffusion length in nanowires suggested that the spin relaxation was suppressed in nanowires compared with bulk materials. This could be possibly attributed to that the phonon scattering is significantly suppressed in nanowires because of a reduced density of states (31), so that the momentum relaxation and hence the spin relaxation is effectively reduced in the Elliott-Yafet mechanism. Also, the one-dimensional channel confines the momentum along the wire axis, and hence minimizes the spin dephasing induced by the intrinsic internal magnetic field in the D'yakonov-Perel' spin relaxation mechanism (32-33). Therefore, the enhancement in the spin lifetime and diffusion length provides another advantage for nanostructures over bulk materials in fabricating practical spintronic devices




Iac+IdcVnonlocal- +V

Ge

SiO2

Si

By Field Fe/MgO

-300 -200 -100 0 100 200 300-12.6

-12.5

-12.4

-12.3

-12.2

-12.1

-12.0

RN

L (

k ΩΩ ΩΩ)

Magnetic Field By (Oe)

T = 40 K

(d)

∆∆∆∆RNL

(a)

(c)

20 nmSiO2

Ge

FeMgO(b)

-0.50 -0.25 0.00 0.25 0.50-4

-2

0

2

4

Cu

rren

t ( µµ µµ

A)

Voltage (V)

T=300 K

T=10 K

tMgO=3.0 nm

0 100 200 300300

400

500

600

700

RA

(µµ µµΩΩ ΩΩ

-cm

2 )

T (K)

Vbias=-0.4 V

Figure 3. (a) Schematics of nonlocal spin valve measurements on the Ge nanowire with Fe/MgO tunnel junctions. (b) Cross-sectional TEM image of a Fe/MgO/Ge nanowire tunnel junction, showing a uniform MgO coverage on the Ge nanowire surface. (c) Temperature-dependent I-V measurements of the Fe/MgO/Ge nanowire device with a MgO thickness of 3.0 nm. The nonlinear I-V characteristics, along with the week temperature dependence of the resistance-area (RA) product plotted in the inset, suggest apparent tunneling nature of the Fe/MgO junction. (d) Nonlocal spin valve signal of Ge nanowires at T = 40 K with an injection ac current of 1 µA, showing a nonlocal resistance of ∆RNL = 470 Ω. The black and red arrows indicate the sweeping direction of the magnetic field, while the blue arrows denote the relative magnetization direction of the spin injector and detector. Electric Field-Controlled Ferromagnetism in MnxGe1-x DMS Nanowires

Furthermore, we successfully demonstrated the pattern-assisted growth of MnxGe1-x

DMS nanowires, which served as the transpinor channel. The growth of the MnxGe1-x nanowires was performed in an ultrahigh-vacuum MBE chamber on a Ge substrate patterned with SiO2 nanotrenches as the mask. The low growth temperature of about 180 °C was selected to avoid the formation of precipitates or secondary phases (34-35). High-purity Ge and Mn sources were evaporated from Knudsen effusion cells. The Ge growth rate was kept at 0.2 Å/s with a tunable Mn flux as the dopant source. The nominal deposition thickness was about 60 nm, and the Mn doping concentration was estimated to




be about x = 2%. After the growth, the surface morphology of the formed MnxGe1-x nanowires was characterized by scanning electron microscopy (SEM), as shown in Figure 4(a). The SEM image clearly shows that the MnxGe1-x nanowires have a very smooth surface with a diameter of about 160 nm. To better illustrate the structure, a schematic diagram of the grown MnxGe1-x nanowires on the Ge substrate is shown in the inset of Figure 4(a).

-4 -2 0 2 4-1.0

-0.5

0.0

0.5

1.0

-0.3 0.0 0.3-0.3

0.0

0.3

400 K

M (µµ µµ

B /

Mn

)

H (kOe)

10 K 50 K 100 K 200 K 300 K 400 K

-80 -40 0 40 80

50

100

150

-250 0 250-0.5

0.0

0.5 70V 40V 0V

M

(µµ µµ

B /

Mn

)

H (Oe)

-40V -70V

HC (

Oe)

VG (V)

500 nm

MnxGe1-x NW

3 nm 500 nm

MnxGe1-x

AuCrSiO2Ge

T = 100 K

(a)

(c) (d)

(b)

Figure 4. (a) SEM image of the grown MnxGe1-x nanowires on a Ge substrate. The inset shows the schematic diagram of the sample structure. (b) Cross-sectional TEM image of the grown MnxGe1-x nanowires on the Ge substrate with the SiO2 pattern. The inset shows the high-resolution TEM images of the MnxGe1-x nanowire. (c) Magnetic hysteresis loops of the MnxGe1-x nanowires at various temperatures. The inset shows the magnified hysteresis at 400 K. (d) Gate voltage-dependent coercivity of the MnxGe1-x nanowires. The inset shows the magnetic hysteresis loops at various gate voltages.

To further understand the microstructure of the grown MnxGe1-x nanowires, the cross-sectional TEM image was taken as shown in Figure 4(b). It can be seen that the MnxGe1-x nanowires were located on the Ge substrate and separated by the SiO2 pattern. The top Cr/Au layers were deposited to prevent ion damages on the MnxGe1-x nanowires during the sample preparation using focused ion beam (FIB). The high-resolution TEM image (shown in the inset) clearly shows the detailed atomic crystal structure of the MnxGe1-x nanowires, which have a perfect epitaxial relation to the underlying Ge (111) substrate. No crystalline defects, such as Mn precipitates or secondary phases, were observed (7). In addition, the magnetic properties of the grown MnxGe1-x nanowires were measured by




superconducting quantum interference device (SQUID) with in-plane magnetic fields. Figure 4(c) shows the temperature-dependent magnetic hysteresis loops at 10 K, 50K, 100 K, 200 K, 300 K and 400 K, respectively. As expected, the coercivity was reduced with the temperature increasing. However, clear hysteresis loop can still be observed even at 400 K as shown in its inset, which provides the direct evidence that the Curie temperature Tc of the MnxGe1-x nanowires is beyond 400 K.

With the single crystallinity and high Tc, it is suggested that our MnxGe1-x nanowire was a perfect DMS system in which the ferromagnetism was mediated by itinerant carriers (36). We then further demonstrated the electric-field controlled ferromagnetism using a metal-oxide-semiconductor capacitor (MOSCAP) structure (7, 37). The semiconductor channel was made of MnxGe1-x nanowires, and the gate dielectric of 60 nm thick Al2O3 was deposited by atomic layer deposition (ALD). By sweeping the gate voltage from positive to negative, the carrier density in the MnxGe1-x DMS nanowires was modulated. Figure 4(d) shows the gate-voltage dependent coercivity, and the inset shows the magnified magnetic hysteresis recorded at various gate voltages. It is found that the coercivity decreased with increasing positive gate voltage, which also depleted holes in the MnxGe1-x nanowire. The reduced hole density weakened the p-d exchange coupling with the magnetic Mn ions (38), leading to the diminishing ferromagnetism. This result confirmed that the ferromagnetism in the MnxGe1-x nanowire was mediated by itinerant holes, and hence it can be controlled by external electric field. It is also observed that the coercivity remained nearly constant with increasing negative gate voltage, which was expected to accumulate holes in the MnxGe1-x nanowire. It is then implied that the hole density at zero gate bias was sufficient to align the magnetic Mn ions inside, and the increased hole density with negative gate voltage did not significantly enhance the ferromagnetism. Similar phenomenon was also observed in our previous DMS system with single-crystalline Mn-doped Ge QDs (7). Our results suggested that the use of DMS nanostructures (quantum dots and nanowires) could enhance the carrier-mediate ferromagnetism and meanwhile minimize crystal defects because of the quantum confinement.

Conclusions

In summary, we discussed spinFETs as a promising candidate for beyond CMOS technologies to address the critical challenges in the Si CMOS scaling. We proposed a nonvolatile transpinor based on MnxGe1-x DMS nanowire with electric field-controlled paramagnetism-to-ferromagnetism phase transition to minimize the charge current flow and reduce the power dissipation. We presented our recent experimental results of electrical spin injection into Ge nanowires using both ferromagnetic Mn5Ge3 Schottky contacts and Fe/MgO tunnel junctions. Much longer spin lifetimes and diffusion lengths were observed in both p- and n-type Ge nanowires compared with bulk Ge. We further demonstrated the pattern-assisted MBE growth of single-crystalline MnxGe1-x DMS nanowires with a Curie temperature above 400 K. The carrier-mediated ferromagnetism was further modulated by an external gate voltage. The discovery of high-Tc MnxGe1-x DMS nanowires with electric-field controlled ferromagnetism, along with the electrical spin injection into Ge nanowires, lays the foundation to build the proposed transpinor device as a possible realization of all-spin logic devices with built-in memory for low-power applications.




Acknowledgments

This work was in part supported by Western Institution of Nanoelectronics (WIN) and National Science Foundation through grant ECCS 1308358. The authors acknowledge experimental assistance and insightful discussions from Device Research Laboratory members. We also thank our collaborators Prof. Lih-Juann Chen, Dr. Chiu-Yen Wang for their significant contributions of TEM analysis, and Prof. R. Kawakami group for their help on the Fe/MgO deposition.

References 1. International Technology Roadmap of Semiconductors (ITRS), http://www.itrs.net

(2012 Edition). 2. I. Žutić, J. Fabian and S. Das Sarma, Rev. Mod. Phys., 76, 323 (2004). 3. S. A. Wolf, D. D. Awschalom, R. A. Buhrman, J. M. Daughton, S. v. Molnar, M.

L. Roukes, A. Y. Chtchelkanova and D. M. Treger, Science, 294, 1488 (2001). 4. S. Datta and B. Das, Appl. Phys. Lett., 56, 665 (1990). 5. S. Sugahara and M. Tanaka, Appl. Phys. Lett., 84, 2307 (2004). 6. B. Behin-Aein, D. Datta, S. Salahuddin and S. Datta, Nature Nanotech., 5, 266

(2010). 7. F. Xiu, Y. Wang, J. Kim, A. Hong, J. Tang, A. P. Jacob, J. Zou and K. L. Wang,

Nature Mater., 9, 337 (2010). 8. D. E. Nikonov and G. I. Bourianoff, IEEE Trans. Nanotech., 4, 206 (2005). 9. S. Salahuddin and S. Datta, Appl. Phys. Lett., 90, 093503 (2007). 10. L. Liu, C.-F. Pai, Y. Li, H. W. Tseng, D. C. Ralph and R. A. Buhrman, Science,

336, 555 (2012). 11. P. K. Amiri and K. L. Wang, Spin, 02, 1240002 (2012). 12. F. J. Jedema, A. T. Filip and B. J. van Wees, Nature, 410, 345 (2001). 13. F. J. Jedema, H. B. Heersche, A. T. Filip, J. J. A. Baselmans and B. J. van Wees,

Nature, 416, 713 (2002). 14. J. Tang, C.-Y. Wang, L.-T. Chang, Y. Fan, T. Nie, M. Chan, W. Jiang, Y.-T.

Chen, H.-J. Yang, H.-Y. Tuan, L.-J. Chen and K. L. Wang, Nano Lett., 13, 4036 (2013).

15. G. Schmidt, D. Ferrand, L. W. Molenkamp, A. T. Filip and B. J. van Wees, Phys. Rev. B, 62, R4790 (2000).

16. A. Fert and H. Jaffre`s, Phys. Rev. B, 64, 184420 (2001). 17. E. I. Rashba, Phys. Rev. B, 62, R16267 (2000). 18. W. Han, Y. Zhou, Y. Wang, Y. Li, J. J. I. Wong, K. Pi, A. G. Swartz, K. M.

McCreary, F. Xiu, K. L. Wang, J. Zou and R. K. Kawakami, J. Cryst. Growth, 312, 44 (2009).

19. K. Yamane, K. Hamaya, Y. Ando, Y. Enomoto, K. Yamamoto, T. Sadoh and M. Miyao, Appl. Phys. Lett., 96, 162104 (2010).

20. S. P. Dash, S. Sharma, J. C. Le Breton, J. Peiro, H. Jaffrès, J. M. George, A. Lemaître and R. Jansen, Phys. Rev. B, 84, 054410 (2011).

21. A. Toriumi, T. Tabata, C. Hyun Lee, T. Nishimura, K. Kita and K. Nagashio, Microelectron. Eng., 86, 1571 (2009).




22. J. Tang, C.-Y. Wang, F. Xiu, A. J. Hong, S. Chen, M. Wang, C. Zeng, H.-J. Yang, H.-Y. Tuan, C.-J. Tsai, L. J. Chen and K. L. Wang, Nanotechnology, 21, 505704 (2010).

23. J. Tang, C.-Y. Wang, F. Xiu, M. Lang, L.-W. Chu, C.-J. Tsai, Y.-L. Chueh, L.-J. Chen and K. L. Wang, ACS Nano, 5, 6008 (2011).

24. J. Tang, C.-Y. Wang, F. Xiu, Y. Zhou, L.-J. Chen and K. L. Wang, Adv. Mater. Sci. Eng., 2011, 316513 (2011).

25. J. Tang, C.-Y. Wang, M.-H. Hung, X. Jiang, L.-T. Chang, L. He, P.-H. Liu, H.-J. Yang, H.-Y. Tuan, L.-J. Chen and K. L. Wang, ACS Nano, 6, 5710 (2012).

26. J. Tang, C.-Y. Wang, W. Jiang, L.-T. Chang, Y. Fan, M. Chan, C. Wu, M.-H. Hung, P.-H. Liu, H.-J. Yang, H.-Y. Tuan, L.-J. Chen and K. L. Wang, Nano Lett., 12, 6372 (2012).

27. M. Julliere, Phys. Lett. A, 54, 225 (1975). 28. S. Iba, H. Saito, A. Spiesser, S. Watanabe, R. Jansen, S. Yuasa and K. Ando, Appl.

Phys. Express, 5, 053004 (2012). 29. Y. Zhou, W. Han, L.-T. Chang, F. Xiu, M. Wang, M. Oehme, I. A. Fischer, J.

Schulze, R. K. Kawakami and K. L. Wang, Phys. Rev. B, 84, 125323 (2011). 30. W. H. Butler, X. G. Zhang, T. C. Schulthess and J. M. MacLaren, Phys. Rev. B, 63,

054416 (2001). 31. A. A. Balandin, Journal of Nanoscience and Nanotechnology, 5, 1015 (2005). 32. A. W. Holleitner, V. Sih, R. C. Myers, A. C. Gossard and D. D. Awschalom, Phys.

ev. Lett., 97, 036805 (2006). 33. A. A. Kiselev and K. W. Kim, PHYSICAL REVIEW B, 61, 13115 (2000). 34. M. Jamet, A. Barski, T. Devillers, V. Poydenot, R. Dujardin, P. Bayle-Guillemaud,

J. Rothman, E. Bellet-Amalric, A. Marty, J. Cibert, R. Mattana and S. Tatarenko, Nat. Mater., 5, 653 (2006).

35. F. Xiu, Y. Wang, K. Wang, Y. Zhou, X. Kou, J. Zou and K. L. Wang, Nanotechnology, 21, 225602 (2010).

36. Y. D. Park, A. T. Hanbicki, S. C. Erwin, C. S. Hellberg, J. M. Sullivan, J. E. Mattson, T. F. Ambrose, A. Wilson, G. Spanos and B. T. Jonker, Science, 295, 651 (2002).

37. H. Ohno, D. Chiba, F. Matsukura, T. Omiya, E. Abe, T. Dietl, Y. Ohno and K. Ohtani, Nature, 408, 944 (2000).

38. Y.-J. Zhao, T. Shishidou and A. J. Freeman, Phys. Rev. Lett., 90, 047204 (2003).




Documents

Spin Transport in Ge Nanowires for Diluted …drl.ee.ucla.edu/wp-content/uploads/2017/07/spin...Spin Transport in Ge Nanowires for Diluted Magnetic Semiconductor-based Nonvolatile