progress report of (bi,sb)2te3 growth - national tsing hua

Chun-Chia Chen (陳俊嘉)Department of Physics, National Tsing-Hua University, Taiwan

Advisor: J. Raynien Kwo (郭瑞年)Department of Physics, National Tsing-Hua University, Taiwan

Advisor: Minghwei Hong (洪銘輝)Department of Physics, National Taiwan University, Taiwan

NTHU/NTU C. C. ChenAdvanced Nano Thin Film Epitaxy Lab 1

Thin Film Growth of Topological Insulators Bi2Se3 and (Bi,Sb)2Te3

toward Bulk-insulating Features and Enhanced Interfacial Exchange Coupling on Rare Earth Iron

Garnets

Acknowledgement

• MBEC. C. Chen and K. H. M. Chen

• SputteringC. C. Tseng, M. X. Guo, and Dr. C. N. Wu

• XRDC. K. Cheng

• TEMDr. C. T. Wu

• TransportS. R. Yang, J. S. Wei, Y. T. Fanchiang, and W. J. Zou

• XPSS. W. Huang and K. Y. Lin

• ARPESS. W. Huang and K. Y. Lin

• FMRY. T. Fanchiang

• Spin pumpingY. T. Fanchiang

• Discussion & suggestionDr. C. N. Wu, Dr. Y. K. Liu, and Y. H. Lin

2

Outline

• Introduction• 3D topological insulators (TIs)• Breaking time reversal symmetry by magnetic proximity effect (MPE)

• Demand for a reproducible growth method • Manipulating the Fermi level (EF) toward the Dirac point• Motivation

• Bi2Se3 / Rare earth iron garnet (ReIG)• Advantage of ReIGs

• Yttrium iron garnet (YIG) & thulium iron garnet (TmIG)• Growth method: Se-buffered low-temperature (SBLT) growth• Improvement in crystallinities (vs. conventional two-step growth)• Enhancement of interfacial exchange coupling

• Observation of anomalous Hall effect (AHE)• Large negative shift in Hres probed by ferromagnetic resonance (FMR)

3

Outline

• (Bi,Sb)2Te3 / 𝛂𝛂-Al2O3 & (Bi,Sb)2Te3 / TmIG• Characterization of structural properties

• Higher-temperature growth of Sb2Te3 & (Bi,Sb)2Te3• Evidences of manipulating EF toward bulk-insulating features

• Detection of two conductive transport channels• Band structure engineering revealed by ARPES

• Electronic transport & magnetoresistance analysis• Observation of room-temperature AHE

• Conclusion

4

3D topological insulators (TIs)Conduction

band

Valence band

Dirac point Γkx

ky

• 3D TIs: Bi2Se3, Bi2Te3, and Sb2Te3• Band inversion induced by strong SOC• Protected by time-reversal symmetry

• Topological surface states (TSSs)• Spin-momentum locking• Massless Dirac fermion

• More…• High spin-to-charge conversion• Weak anti-localization (WAL) effect• Forbidden-back-scattering transport

Spin-momentum locked

No Spin-momentum lockingM. Brahlek et al., Solid State Commun. 215-216, 54-62 (2015).5

Breaking Time Reversal Symmetry (TRS) in TIsQuantum anomalous Hall effect (QAHE)

Breaking TRS by MPE

6

Features• Quantized plateau ( 𝒉𝒉

𝒆𝒆𝟐𝟐) in 𝝆𝝆𝒙𝒙𝒙𝒙

• Approaching zero resistivity in 𝝆𝝆𝒙𝒙𝒙𝒙• Dissipationless transport

C.-Z. Chang et al., Science 340(6129), 167-170 (2013).C.-Z. Chang et al., Nat. Mater. 14, 473–477 (2015).C.-Z. Chang et al., J. Phys.: Condens. Matter 28, 123002 (2016).

Cr-doped (Bi,Sb)2Te3

V-doped (Bi,Sb)2Te3

Transition metal doping

C.-Z. Chang et al., Science 340 (6129), 167-170 (2013).

QAHE in Cr-doped TI Below 2 K

Magnetic proximity effectAdvantages of MPE:• Higher TC than that of

magnetic doping• No crystal defects• Uniform magnetization

C. Tang et al., Sci. Adv. 3 (6), e1700307 (2017).

Breaking Time Reversal Symmetry (TRS) in TIs

Ways to break TRS in TIs: Introducing magnetic moments


Magnetic proximity effect (MPE)

2010 2012 2014 20160

100

200

300

400 MPE Doping MPE Doping+MPE

T C (K)

Year

Doping

𝑬𝑬

= ± ℏ𝒗𝒗𝑭𝑭𝒌𝒌𝒙𝒙 +𝑱𝑱𝟐𝟐𝑴𝑴𝒙𝒙

𝟐𝟐

+ ℏ𝒗𝒗𝑭𝑭𝒌𝒌𝒙𝒙 −𝑱𝑱𝟐𝟐𝑴𝑴𝒙𝒙

𝟐𝟐

+𝑱𝑱𝟐𝟐𝑴𝑴𝒛𝒛

𝟐𝟐

H. Z. Lu et al., Phys. Rev. Lett. 107, 076801 (2011).

Quantum anomalous Hall effect (QAHE)

Breaking TRS by MPE

6








Breaking TRS by MPE

• Persistence of MPE to room temperature • Low TC (~17 K)• In-plane magnetic anisotropy (IMA) of EuS

TI/MI hetero-structure with a very small lattice mismatch

EuSBi2Se3

F. Katmis et al., Nature 533, 513 (2016).

6




Tensile-strained Tm3Fe5O12(thulium iron garnet, TmIG)

• Magnetic insulator• High TC (above 500 K)• Perpendicular magnetic anisotropy (PMA)C. N. Wu et al., AIP Adv. 8 (5), 055904 (2018).C. N. Wu et al., Sci. Rep. 8 (1), 11087 (2018).

TmIG





Breaking TRS by MPE

6




Tensile-strained Tm3Fe5O12(thulium iron garnet, TmIG)

• Magnetic insulator• High TC (above 500 K)• Perpendicular magnetic anisotropy (PMA)C. N. Wu et al., AIP Adv. 8 (5), 055904 (2018).C. N. Wu et al., Sci. Rep. 8 (1), 11087 (2018).

TmIG


Breaking TRS by MPE

However,• Complicated surface atomic

arrangement

• Huge lattice mismatch with Bi2Se3

6

Conventional two-step growth leads to:

1st step: Low temp.

2nd step: High temp.

Ordered and smooth initial layer

Atoms with more kinetic energy Better film quality

N. Bansal et al., Thin Solid Films 520 (1), 224 (2011).

110

230

Conventional two-step growth

Applied on rare earth iron garnets (ReIGs):Atoms tend to bond to the surface dangling bonds first

Breaking TRS by MPE

7

Z. Jiang et al., AIP Adv. 6 (5), 055809 (2016).

H. Wang et al., Phys. Rev. Lett. 117 (7), 076601 (2016).

YIG

Interlayer with poor crystallinity

Conventional two-step growth leads to:Large variation of TC

of MPE properties

TC ~ 100 K

Z. Jiang et al., Nano Lett. 15 (9), 5835 (2015).

(BixSb1-x)2Te3

interlayer

interlayer

TC ~ 20 - 150 KR AH

E(Ω

)

Temperature (K)

Tem

pera

ture

(K)

Bi ratio (x)

Breaking TRS by MPE

Z. Jiang et al., Nat. Commun. 7, 11458 (2016).

Urgent demand for a reproducible growth method • Narrowing the variation of film qualities• High-quality interface• Strong exchange coupling

7

Manipulating EF toward Dirac point

Bulk conduction band+

Surface states


Surface states

Bulk valence band+

Surface states• Bulk states dominate the transport properties.• TSS features are diluted by the bulk contribution.• It is hard to distinguish the contributions between

the TSS and the bulk.

Bi2Se3 Bi2Te3 Sb2Te3

G. Wang et al., Nano Res. 3(12),874–880 (2010).

Y.-Y. Li et al., Adv. Mater. 22,4002–4007 (2010).

Y. Zhang et al., Nat. Phys. 6,584-588 (2010).

8

Manipulating EF toward Dirac pointWays to eliminate/reduce the

bulk contribution in TIs

Electric gating effect

J. S. Lee et al., npj Quantum Mater. 3, 51 (2018).

M. Lang et al., Nano Lett. 13, 48−53 (2013).

Electric gating effect• Tuning EF by the field effect• Affecting only one side of the TI• No modification on the band structure

9

Manipulating EF toward Dirac pointWays to eliminate/reduce the

bulk contribution in TIs

Electric gating effect

Chemical doping / alloy

• Tuning the position of the EF• Introducing different amounts

of defects• Modifying the band structure

of TI

Bi2Te3TeBi anti-site

n-type doping

Sb2Te3SbTe anti-site

p-type doping

J. M. Zhang et al. Phys. Rev. B 88, 235131 (2013)

x = 0.0 x = 0.25 x = 0.62 x = 0.75

x = 0.88 x = 0.94 x = 0.96 x = 1.0J. Zhang, et al., Nat. Commun. 2, 574 (2011).

(Bi,Sb)2Te3 (BST)

Manipulating the EF into the bulk band gap bulk insulating TI 9

MotivationMolecular Beam Epitaxy (MBE)

Bi2Se3

𝛂𝛂-Al2O3

BiSe

Reflection High-Energy Electron Diffraction

(RHEED)

Base pressure:4×10-10 torr

(2×10-10 torr)

10

2D diffractionStreaks: smooth surface & Great crystallinityRings or Spots: polycrystalline or rough surface

Pure element sourceCoiled with filament

Motivation

Bi2Se3𝛂𝛂-Al2O3

Bi2Se3ReIG

• Adopting a new growth method (SBLT growth)

• Improvement of interface quality

(Bi,Sb)2Te3𝛂𝛂-Al2O3

• Toward bulk-insulting feature

• Exhibition of pure surface state

(Bi,Sb)2Te3TmIG

• Breaking time reversal symmetry

• Pushing the TC up to room temperature

12

H. Zhang et al., Nat. Phys. 5 (6), 438 (2009).

BiSe1Se2

Se1BiSe2BiSe1

Se1

Se1Quintuple layer (QL)

(~ 1 nm)

I: Bi2Se3/TmIG&

Bi2Se3/YIG

2NTHU/NTU C. C. ChenAdvanced Nano Thin Film Epitaxy Lab13

• Bi2Se3 / Rare earth iron garnet (ReIG)• Advantage of ReIGs• Growth method: Se-buffered low-temperature (SBLT) growth• Improvement in crystallinities (vs. conventional two-step growth)• Enhancement of interfacial exchange coupling

• Thulium iron garnet (Tm3Fe5O12, TmIG)• Perpendicular magnetic anisotropy (PMA) (induced by tensile strain)

• Yttrium iron garnet (Y3Fe5O12, YIG)• In-plane magnetic anisotropy (IMA) and low damping constant

ReIG substrate preparation

Garnet structure

Sputter MBEWithout chemical cleaning

150 15 minOutgassing

Growth temp.

at 500Bi2Se3/YIG

at 600, 1hr(Bi,Sb)2Te3/YIG

(Bi,Sb)2Te3/TmIG at 600, 1hr

H. Wang et al., Phys. Rev. Lett. 117, 076601 (2016).

Z. Jiang et al., AIP Adv. 6, 055809 (2016).


Outgassing

1𝟓𝟓𝟓𝟓Clear line

4𝟓𝟓𝟓𝟓Rings

RHEED patterns:350 ring (polycrystalline)

Residual gas analysis:300 CxOx compounds

14

Demand for a reproducible growth method

However,• Complicated surface atomic

arrangement

• Huge lattice mismatch with Bi2Se3

Garnet structure

1st step: Low temp.

2nd step: High temp.

Ordered and smooth initial layer

Atoms with more kinetic energy

Better film qualityN. Bansal et al., Thin Solid Films 520 (1), 224 (2011).

110

230

Conventional two-step growth

Z. Jiang et al., AIP Adv. 6 (5), 055809 (2016).

Interlayer with poor crystallinity

interlayer

YIG

(BixSb1-x)2Te3

Urgent demand for a reproducible growth method • Narrowing the variation of film

qualities• High-quality interface• Strong exchange coupling

15

The Se-buffered low-temperature (SBLT) growth:Bi2Se3

Amorphous Se

TmIG

TmIG(111) (~50) amorphous

The substrate wascovered by a layerof amorphous Seand ~1 nm BixSe1-x.

~50

The SBLT growth

16

amorphous ~1 QL Bi2Se3

The Se-buffered low-temperature (SBLT) growth:

Bi2Se3

Amorphous Se

TmIG

The high-quality Bi2Se3at the 1st QL serves as agreat template for thefurther Bi2Se3 growth.

250

The SBLT growth

16

~1 QL Bi2Se3

The Se-buffered low-temperature (SBLT) growth:

Bi2Se3

Amorphous Se

TmIG

Sharper and brighterRHEED patterns wereobtained in thickerBi2Se3 films.~7 QL Bi2Se3

250

The SBLT growth

16

Bi2Se3 (~1 QL) Bi2Se3 (~3 QL) Bi2Se3 (~7 QL)


TmIG

SBLT growth

Two-step growth

Spotty ring featurepoly crystalline

Line featurebetter crystallinity

and smooth surface

Grain growth,forming 2D surface

Reaching a good-quality surface

More distinct RHEED patternsachieving excellent

crystallinity of thin films

Comparison of RHEED patterns

17

YIG Bi2Se3 (~1 QL) Bi2Se3 (~1 QL) Bi2Se3 (~1 QL)

It works on Bi2Se3/YIG !



TmIG

SBLT growth

Two-step growth

Comparison of RHEED patterns

17

15

-1

0

1

2

nm

-5

5

0 1 2 3-6-3036

Heig

ht (n

m)

Position (µm)

Roughness = 2.58 nm

0 1 2 3-6-3036

Heig

ht (n

m)

Position (µm)

Roughness = 0.43 nm

1

-1

0

nm

200 𝒏𝒏𝒎𝒎Roughness = 0.47 nm

0 500 1000-2-101

Heig

ht (n

m)

Position (nm)

Step height = ~1nm

Top surface:• Smooth surface• Clear triangular

domain

Two-step growth SBLT growth

600 𝒏𝒏𝒎𝒎 600 𝒏𝒏𝒎𝒎

nm

GGGTmIG

Bi2Se37nm

Structural analyses

18

15

-1

0

1

2

nmnm

-5

0 1 2 3-6-3036

Heig

ht (n

m)

Position (µm)

600 𝒏𝒏𝒎𝒎

5

600 𝒏𝒏𝒎𝒎Roughness = 2.58 nm

0 1 2 3-6-3036

Heig

ht (n

m)

Position (µm)

Roughness = 0.43 nm

• Smooth surface• Clear triangular

domain

0 500 1000012

Heig

ht (n

m)

Position (nm)Roughness = 1.29 nm Roughness = 0.58 nm

Roughness = 0.47 nmGGGYIG

Bi2Se37nm

nm

-4

12

4

nm

-4

4

0

-81 𝝁𝝁𝒎𝒎 1 𝝁𝝁𝒎𝒎


GGGTmIG

Bi2Se37nm


Top surface:

Structural analyses

18

15

-1

0

1

2

nmnm

-5600 𝒏𝒏𝒎𝒎

5

600 𝒏𝒏𝒎𝒎Roughness = 2.58 nm Roughness = 0.43 nm

• Smooth surface• Clear triangular

domain GGGTmIG

Bi2Se37nm

0 500 1000012

Heig

ht (n

m)

Position (nm)Roughness = 1.29 nm Roughness = 0.58 nm

Roughness = 0.47 nm

nm

-4

12

4

nm

-4

4

0

-81 𝝁𝝁𝒎𝒎 1 𝝁𝝁𝒎𝒎

200 𝒏𝒏𝒎𝒎GGGYIG

Bi2Se37nm


Roughness = 0.71 nmH. Wang et al., Phys. Rev. Lett. 117 (7), 076601 (2016).

Roughness = 0.77 nmC. Tang et al., Sci. Adv. 4, eaas8660 (2018).


1 𝝁𝝁𝒎𝒎

(Bi0.3Sb0.7)2Te3/YIG

Bi2Se3/YIG

Top surface:

Structural analyses

18

Bi2Se3

TmIG5 nm

Bi2Se3

TmIG

0.5 nm

5 nm

HRTEM images

0.5 nm

Cs-corrected STEM images

Interface: ~1nm amorphous

interlayer

Films: Clear atomic

structuresInterface: Se buffer layer

evaporated Visible atoms in

the 1st QL

Two-step growth

SBLT growth

Bi2Se3 (~1 QL)

Structural analyses

19

168 164 160 156 1520

5

10

15

20

Inte

nsity

(kcp

s)

Binding Energy (eV)

(ele

men

tal S

e)Se

3p 1/

2Se

3p 1/

2

Bi 4f5/2

Se 3

p 3/2

(Bi 2Se

3)Se

3p 3/

2

Bi 4f7/2

32 30 28 26 24 220

1

2

3 Bi 5d

Inte

nsity

(kcp

s)

Binding Energy (eV)

740 730 720 7103.8

4.0

4.2

4.4Fe 2p

Inte

nsity

(kcp

s)

Binding Energy (eV)

60 58 56 54 52 50 480

1

2Se 3d5/2

(Bi2Se3)

Se 3d3/2

Se 3d5/2(elemental Se)

Se 3d

Inte

nsity

(kcp

s)

Binding Energy (eV)

XPS spectra of 4 nm Bi2Se3 on TmIG

No elemental Se was detected at the interface!

GGGTmIG

Bi2Se3 4 nm

20

Interfacial analyses

-300 0 300

180 K

140 K

100 K10 mΩ

R AHE (

mΩ

)

B (Gs)

70 K

220 K

Anomalous Hall effect (AHE) loops

Magnetic proximity effect (MPE):

Spin current effect at the interface:


Y. Lv et al., Nat. Commun. 9 (1), 111 (2018).

Bulk state:Spin Hall effect

Surface states:Rashba-Edelstein effect

+

Magnetization ofthe bottom

topological surface state (TSS)

Enhanced exchange coupling at the interface!

Spin current effect has been precluded in our other work

S. R. Yang et al., arXiv: 1811.00689 (2018).(Accepted by Phys. Rev. B)

Observation of AHEGGGTmIGBi2Se3

21

0 6 12 18 24 30 36 42-0.16

-0.12

-0.08

-0.04

0.00

0.04TSS-induced interfacial anisotropy

K i (erg

/cm

2 )dBS (nm)

3600 4000 4400norm

aliz

ed d

I/dH(

a.u.

)

H (Oe)

365 Oe

TmIG(15) Bi2Se3(7)/TmIG(15)

4000 4400 4800norm

aliz

ed d

I/dH(

a.u.

)

H (Oe)

137 Oe

TmIG(15) Bi2Se3(7)/TmIG(15)

Two-step growth

SBLT growth Y. T. Fanchiang et al., Nat. Commun. 9, 223 (2018)

Our previous work

Ferromagnetic resonance

(FMR)

Spin dynamics at the interface

Negative shift in Hres of TmIG

Enhancement of IMA at the interface

Stronger interfacial exchange coupling !

Bi2Se3 grown on TmIG

Enhanced interfacial exchange coupling

22

• Bright and distinct RHEED at the 1st QL• Smoother surface and sharp triangular domains• High-quality interface• Observation of AHE• Larger shift in Hres

Our SBLT growth attained:

Enhanced interfacial exchange coupling

Summary I

23C. C. Chen et al., Appl. Phys. Lett. 114, 031601 (2019).(Featured, cover story)

II: (Bi,Sb)2Te3/𝜶𝜶-Al2O3&

(Bi,Sb)2Te3/TmIG


• (Bi,Sb)2Te3 / 𝛂𝛂-Al2O3 & (Bi,Sb)2Te3 / TmIG• Characterization of structural properties• Evidences of manipulating EF toward bulk-insulating features• Electronic transport & magnetoresistance analysis

Manipulating EF toward Dirac point


Surface states

• Bulk states dominate the transport properties.

• It is hard to distinguish the contributions between the TSS and the bulk.

Bi2Se3 (our work)K//(Å-1)

E-E F(e

V)

M- Г -M

EF

Bi2Te3TeBi anti-site

n-type doping

Sb2Te3SbTe anti-site

p-type doping

J. M. Zhang et al. Phys. Rev. B 88, 235131 (2013)

• Tuning the position of the EF• Introducing different amounts of defects• Modifying the band structure of TI

Ways to eliminate/reduce the bulk contribution in TIs: (Bi,Sb)2Te3

Manipulating the EF into the bulk band gap bulk insulating TI 25

Sb(x) 10 0.710.500.25

1 QL 1 QL 1 QL 1 QL 1 QL~15 QL ~15 QL ~15 QL ~15 QL ~15 QL

• Streaky RHEED patterns from the very first quintuple layer (QL) to ~15 QL grown by MBE

Excellent crystallinities of BST films

26

Sb(x) 10 0.50

~15 QL ~15 QL

0.710.25

~15 QL ~15 QL ~15 QL

0 10 20 30

54

3

Inte

nsity

(a. u

.)

Time (min)

1QL 2• Streaky RHEED patterns from the very first QL to ~15 QL grown by MBE

• Clear RHEED oscillations indicating the layer-by-layer growth

• Large domains with sharp triangular shapes

10.50

Sb(x) = 0.50

Roughness = 0.86 nm

nm

1

2

0

-2

-1


Excellent crystallinities of BST films

26

0.0 0.5 1.04.26

4.29

4.32

4.35

4.38

In-p

lane

latti

ce c

onst

ant

(a) (

A)

Sb(x) determined by the flux gauge

Sb2Te3a = 4.26 Å

Bi2Te3a = 4.38 Å

XRD (ex-situ) XPS (in-situ)

• Compositions of (Bi,Sb)2Te3 can be both in-situ and ex-situ calibrated by X-ray photoelectron spectroscopy (XPS) and X-ray diffraction (XRD), respectively.

4.26 4.29 4.32 4.35 4.380.0

0.2

0.4

0.6

0.8

1.0

Sb(x

) (ca

libra

ted

by X

PS)

in-plane lattice constant(A) (measured by XRD)

Composition characterization

270.0 0.5 1.029.8

30.0

30.2

30.4

30.6

30.8

Nor

mal

latti

ce c

onst

ant (

c) (A

)

Sb(x)

Sb2Te3c = 30.43 Å

Bi2Te3c = 30.49 Å

Metal-like

Semiconductor-like

Te capping layer & R-T curves


~20 nm Te

100 150 200 250 3000

200

400

600

R shee

t(kΩ

)

T(K)

Our work

At 300 Kn2D = 3.37×1012 cm-2

Rsheet = 45483 ΩMobility = 39.8 cm2/Vs

~ 600 kΩat 100 K

𝛂𝛂-Al2O3

~20 nm Te

To prevent the oxidation of TIs

• Resistance increasing with the temperature decreasing

• Higher value of resistance

• Resistance decreasing with the temperature decreasing

0 50 100 150 200 250 3000.0

0.5

1.0

1.5

2.0

2.5

3.0

R shee

t(kΩ

)

T(K)

(Bi0.29Sb0.71)2Te3

Bi2Te3

(Bi0.5Sb0.5)2Te3

Sb2Te3

28

Metal-like

Semiconductor-like


• Resistance increasing with the temperature decreasing

• Higher value of resistance


0 50 100 150 200 250 3000.0

0.5

1.0

1.5

2.0

2.5

3.0

R shee

t(kΩ

)

T(K)

(Bi0.29Sb0.71)2Te3

Bi2Te3

(Bi0.5Sb0.5)2Te3

Sb2Te3

J. Zhang et al., Nat. Commun. 2, 574 (2011).

28


Metal-like

Semiconductor-like• Resistance increasing with the

temperature decreasing• Higher value of resistance


0 50 100 150 200 250 3000.0

0.5

1.0

1.5

2.0

2.5

3.0

R shee

t(kΩ

)

T(K)

(Bi0.29Sb0.71)2Te3

Bi2Te3

(Bi0.5Sb0.5)2Te3

Sb2Te3

0 50 100 150 200 250 300

2.2

2.3

2.4

2.5

2.6

2.7

2.8

R shee

t(kΩ

)

T(K)

𝑮𝑮𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕 =𝟏𝟏

𝑹𝑹𝟓𝟓 + 𝑨𝑨𝑨𝑨+

𝟏𝟏

𝑹𝑹𝟏𝟏𝐞𝐞𝐞𝐞𝐞𝐞( ∆𝒌𝒌𝑨𝑨)

𝐑𝐑𝟓𝟓 = 𝟐𝟐𝟏𝟏𝟐𝟐𝟐𝟐 𝛀𝛀𝐀𝐀 = 𝟒𝟒.𝟐𝟐 𝛀𝛀/𝐓𝐓𝐑𝐑𝟏𝟏 = 𝟐𝟐𝟓𝟓𝟏𝟏 𝛀𝛀∆=59 meV activation energy(Bi0.29Sb0.71)2Te3

28


0 50 100 150 200 250 300

2.2

2.3

2.4

2.5

2.6

2.7

2.8

R shee

t(kΩ

)

T(K)

𝑮𝑮𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕 =𝟏𝟏

𝑹𝑹𝟓𝟓 + 𝑨𝑨𝑨𝑨+

𝟏𝟏

𝑹𝑹𝟏𝟏𝐞𝐞𝐞𝐞𝐞𝐞( ∆𝒌𝒌𝑨𝑨)

𝐑𝐑𝟓𝟓 = 𝟐𝟐𝟏𝟏𝟐𝟐𝟐𝟐 𝛀𝛀𝐀𝐀 = 𝟒𝟒.𝟐𝟐 𝛀𝛀/𝐓𝐓𝐑𝐑𝟏𝟏 = 𝟐𝟐𝟓𝟓𝟏𝟏 𝛀𝛀∆=59 meV activation energy(Bi0.29Sb0.71)2Te3

B. Skinner et al., Phys. Rev. Lett. 109, 176801 (2012).

• Network of p-n junction

• Activation energy ≈ energy from EFto Ee (effective Ec)

• Larger ∆ of a TI => better quality

• The largest achievable ∆ of doped TI is 𝟓𝟓.𝟏𝟏𝟓𝟓𝐄𝐄𝐠𝐠 ≈ 𝟓𝟓𝟓𝟓𝟓𝟓𝐞𝐞𝟓𝟓 according to simulations.

28

(n-type)(p-type)

Very low n2D & boosted mobility

0.0 0.2 0.4 0.6 0.8 1.0-6

-4

-2

0

2

4

n 2D (x

1013

cm-2)(2

K)(1

5 nm

)

Sb(x)

(n-type)(p-type)

0.0 0.2 0.4 0.6 0.8 1.0

-4

0

4

8

12

16

20

without capping Te capping (LT)Te capping (HT)

n 2D (x

1013

cm-2)(2

K)(5

nm

)

~15 nm (2K)

~5 nm (2K)

n2D:• Demonstration from n-type to

p-type• The lowest value of − 𝟏𝟏.𝟐𝟐 × 𝟏𝟏𝟓𝟓𝟏𝟏𝟐𝟐 𝒄𝒄𝒎𝒎−𝟐𝟐 (15 nm)and −𝟏𝟏.𝟓𝟓 × 𝟏𝟏𝟓𝟓𝟏𝟏𝟐𝟐 𝒄𝒄𝒎𝒎−𝟐𝟐 (5 nm)

Conduction band

Valence band

EF

29

Very low n2D & boosted mobility

0.0 0.2 0.4 0.6 0.8 1.00

400

800

1200

1600

Mob

ility

(cm

2 /Vs)

(2K)

(15

nm)

Sb(x)

0.0 0.2 0.4 0.6 0.8 1.00

200

400

600

800

1000

without capping Te capping (LT)Te capping (HT)M

obili

ty(c

m2 /V

s)(2

K)(5

nm

)

~15 nm (2K)

~5 nm (2K)

Mobility:• Enhanced value with

capping layers• Boosted value with high-

temp. growth • Higher value at Dirac point

stemmed from the feature of Dirac fermion

J. Zhang et al., Nat. Commun. 2, 574 (2011).

Q.-K. Xue’s group

L. He et al., Sci. Rep. 3, 3406 (2013).

K. L. Wang’s group

5 nm (Bi,Sb)2Te3 𝜇𝜇 ~ 500 cm2/Vs

8 nm (Bi,Sb)2Te3 𝜇𝜇 ~ 300 cm2/Vs

(before applying field effect)

Our group5 nm (Bi,Sb)2Te3 𝜇𝜇 = 917 cm2/Vs

30

0.0 0.2 0.4 0.6 0.8 1.0

-1.0

-0.5

0.0

α (1

5 nm

)

Sb(x)

𝜶𝜶~ − 𝟓𝟓.𝟓𝟓: one channel with weak anti-localization (WAL)𝜶𝜶~ − 𝟏𝟏: two channel with WALl : phase coherence length

surface state

surface state surface state

surface state

Bulk insulating

Bulk conductive

∆𝐺𝐺xx = 𝜶𝜶𝑒𝑒2

𝜋𝜋ℎ 𝜓𝜓ℏ

4𝑒𝑒𝑙𝑙2𝐵𝐵+

12

− lnℏ

4𝑒𝑒𝒕𝒕2𝐵𝐵+ 𝑐𝑐𝐵𝐵2

HLN equation:

(1) the Sb composition dependence

(3) the BST thickness dependence(2) the effect of the capping layer

Two conductive transport channels

w/o capping (15 nm)with capping (15 nm) (LT)with capping (15 nm) (HT)w/o capping (5 nm)with capping (5 nm) (LT)with capping (5 nm) (HT)

31

Two conductive transport channels

0.0 0.2 0.4 0.6 0.8 1.00

100

200

300

400

500

600 without capping (15 nm)Te capping (15 nm)(LT)Te capping (15 nm)(HT)

Phas

e co

here

nce

leng

th (n

m)(1

5 nm

)

Sb(x)0.0 0.2 0.4 0.6 0.8 1.00

100

200

300

400

500

600 without capping (5 nm)Te capping (5 nm)(LT)Te capping (5 nm)(HT)

Phas

e co

here

nece

leng

th (n

m)(5

nm

)

Sb(x)

Low phase coherence length of (Bi,Sb)2Te3• Network of p-n junction• Transport by means of hopping or

tunneling between charge puddles• Enhanced coherence length with the

high-temperature growth B. Skinner et al., Phys. Rev. Lett. 109, 176801 (2012).

32

Sb(x) 10 0.780.50

Condition:• He I hν = 21.2 eV• At 300K

• Thickness = ~20 nm• K− Γ− K

• Tuning EF across the Dirac point• Clear surface states shown in all samples

-0.2 -0.1 0.0 0.20.1k// (Å-1)

-0.2 -0.1 0.0 0.20.1

0.0

-0.4

-0.2

0.0

-0.4

-0.2

-0.2 -0.1 0.0 0.20.1

0.0

-0.4

-0.2

-0.2 -0.1 0.0 0.20.1

0.0

-0.4

-0.2

E-EF (eV) Bi2Te3 (Bi0.5Sb0.5)2Te3 Sb2Te3(Bi0.22Sb0.78)2Te3

Tuning EF and band structures

33

Low-temperature growth (modified from Se-buffered low-temp. (SBLT) growth)

(Bi, Sb)2Te3

GGGTmIG

Te

~ 15 nm (220)

TmIG (~ 40)

H.-Z. Lu et al., PRL 107, 076801 (2011).

nm

0

2.5

5

7.5

10

-2.5

Roughness = 1.42 nm


Breaking TRS by MPE

Lower flux ratio

Bi,Sbopen

Low temp.

Bi,Sb,Teclose

Teopen

Bi,Sbopen

Bi,Sbclose

Teopen

High temp.(Initial layer)

Higher flux ratio

34

Modified SBLT growth (Te-based TI)

-800 -400 0 400 800-0.4

-0.2

0.0

0.2

0.4

R AHE (

Ω)

H (Oe)

T = 100K

0 100 200 3000.00

0.05

0.10

0.15

0.20

(Bi,Sb)2Te3 15 nm/TmIG (Bi,Sb)2Te3 5 nm/TmIG

R AHE (

Ω)

T(K)

(Bi, Sb)2Te3

GGGTmIG

Room-temperature AHE was observed!

-800 -400 0 400 800-0.4

-0.2

0.0

0.2

0.4

R AHE (

Ω)

H (Oe)

T = 220K

-2000 -1000 0 1000 2000-0.4

-0.2

0.0

0.2

0.4

R AHE (

Ω)

H (Oe)

T = 30K

-4000 -2000 0 2000 4000-0.15-0.10-0.050.000.050.100.15

R AHE (

Ω)

H (Oe)

T = 300K

Bi2Se3

GGGTmIG

-150 -100 -50 0 50 100 150-30

-15

0

15

30

T = 100K

R AHE (

mΩ

)

H (Oe)-150 -100 -50 0 50 100 150

-400

-200

0

200

400

R AHE (

mΩ

)

H (Oe)

T = 100K

(Bi, Sb)2Te3

GGGTmIG

(Bi,Sb)2Te3/TmIG (in this work) Bi2Se3/TmIG (our previous work)

~ one order larger RAHE than that in Bi2Se3/TmIG

Observation of AHE up to RT

35

(Bi0.2Sb0.8)2Te3 5 nm/TmIG(Bi0.3Sb0.7)2Te3 5 nm/TmIG

(Bi, Sb)2Te3

SGGGTmIG

(J. Shi’s group)

0 100 200 3000.00

0.05

0.10

0.15

0.20

(Bi,Sb)2Te3 15 nm/TmIG (Bi,Sb)2Te3 5 nm/TmIG

R AHE (

Ω)

T(K)

(our group)

(Bi, Sb)2Te3

GGGTmIG


Observation of AHE up to RT

Factors affect the extent of RAH:

• Stronger PMA induced by SGGG• Larger tensile strain

• EF closer to Dirac pointM. Li et al., Phys. Rev. B 96, 201301(R) (2017).

• Interface quality

-4 -2 0 2 4-100

-50

0

50

100

R yx (Ω

)

B (T)

10 K 30 K 50 K 70 K 100 K 140 K 180 K 220 K

close to Dirac point

36

R-T curve

(Bi, Sb)2Te3

GGGTmIG

Te

(Bi, Sb)2Te3

GGGTmIG

Te

~ 15 nm

~ 5 nm

0 100 200 300

2.73.03.33.63.94.2

R s (kΩ

)

T (K)

0 100 200 3007

8

9

10

11

12

R s (kΩ

)

T (K)

• Semiconductor-like feature in both samples

• Similar to the previous work

(Bi0.2Sb0.8)2Te3 5 nm/TmIG


R-T curves of (Bi,Sb)2Te3/TmIG

37

MR

-9 -6 -3 0 3 6 94.0

4.4

4.8

5.2

R s (kΩ

)

B (T)

2 K

-9 -6 -3 0 3 6 9

12

13

14

15

16 2K

R s (kΩ

)

B (T)

(Bi, Sb)2Te3

GGGTmIG

Te

(Bi, Sb)2Te3

GGGTmIG

Te

~ 15 nm

~ 5 nm

MR analysis by HLN equationBi2Se3/TmIG (our work)

Negative MR• Magnetization gap

(our case)• Strong localization (𝑹𝑹𝒔𝒔 ≥ 𝒉𝒉/𝒆𝒆𝟐𝟐)

J. Liao et al., Phys. Rev. Lett. 114, 216601 (2015).

• Hybridization gapH. Z. Lu et al., Phys. Rev. Lett. 107, 076801 (2011).

No negative MR S. R. Yang et al., arXiv: 1811.00689 (2018).

38

MR analysis by HLN equationBi2Se3/TmIG (our work)• No negative MR

• Smaller magnitude of α in (Bi,Sb)2Te3/TmIG than those on Sapphire

• Suppressed WAL

-9 -6 -3 0 3 6 9

1.0

1.2

1.4

R s(B

)/Rs(

0)

B (T)

(Bi,Sb)2Te3 15 nm/sapphire(Bi,Sb)2Te3 15 nm/TmIG

(Bi,Sb)2Te3/TmIG (this work)

0.2 0.3 0.4 0.5 0.6-1.0

-0.8

-0.6

-0.4

-0.2

0.0

BST 5 nm/TmIG BST 15 nm/TmIG BST 5 nm/Sapphire BST 15 nm/Sapphire

α

Fitting Range (T)

38

When the top surface and the bottom surface are decoupled,

∆𝝈𝝈𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕 =𝝈𝝈𝒕𝒕𝒕𝒕𝒕𝒕 𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒕𝒕𝒄𝒄𝒆𝒆 + 𝝈𝝈𝒃𝒃𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒎𝒎 𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒕𝒕𝒄𝒄𝒆𝒆

𝝈𝝈𝒕𝒕𝒕𝒕𝒕𝒕 𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒕𝒕𝒄𝒄𝒆𝒆 = 𝐖𝐖𝐀𝐀𝐖𝐖 (𝜶𝜶 = −𝟓𝟓.𝟓𝟓)

𝝈𝝈𝒃𝒃𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒕𝒎𝒎 𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒔𝒕𝒕𝒄𝒄𝒆𝒆 = 𝐖𝐖𝐀𝐀𝐖𝐖 −𝟓𝟓.𝟓𝟓 < 𝜶𝜶 < 𝟓𝟓 +

surface state

surface state

Bulk insulating

𝐖𝐖𝐖𝐖 (𝟓𝟓 < 𝜶𝜶 < 𝟓𝟓.𝟓𝟓)

MR analysis by HLN equation

0.2 0.3 0.4 0.5 0.6-1.0

-0.8

-0.6

-0.4

-0.2

0.0

BST 5 nm/TmIG BST 15 nm/TmIG BST 5 nm/Sapphire BST 15 nm/Sapphire

αFitting Range (T)

39

Excellent crystallinity in (Bi,Sb)2Te3 on 𝜶𝜶-Al2O3

• Transport: lowest n2D, two conducting channels

• ARPES: closer to Dirac point

surface state

surface state

Bulk insulating

-4000 -2000 0 2000 4000-0.15-0.10-0.050.000.050.100.15

R AHE (

Ω)

B (Oe)

T = 300K

surface state

surface state

Bulk insulating

40

Excellent crystallinity in (Bi,Sb)2Te3 on TmIG

• Room-temperature AHE

• Weak thickness dependence insulating bulk

• No negative MR with smaller magnitude of 𝛼𝛼

MPE only affected the bottom surface

Summary II

Conclusion


Conclusion

Bi2Se3ReIG


(Bi,Sb)2Te3TmIG

• Breaking time reversal symmetry• Improvement of the interface quality• Enhancement of the interfacial exchange coupling

• Toward insulating bulk state• Two conductive transport channels• Band engineering revealed by ARPES

• Modifying the growth from Bi2Se3/ReIG• Pushing the TC up to room temperature (AHE)

42

progress report of (bi,sb)2te3 growth - national tsing hua

Documents