arXiv:1701.00582v1 [cond-mat.mes-hall] 3 Jan 2017 h inequivalent the . gradient, oe in posed rents. ehnssvaeetia en.Rfrne[7 sug- [27] Reference the that means. gested to electrical similar via coupling cou- mechanisms valley-orbit valley-processing inter-valley promising as thus providing coupling act and -orbit that point Γ arises same pling the into folded eeti h td fvletoisdet h presence the to due valleys valleytronics inequivalent of in- two study of special two-dimensional the of In in to are terest controllable systems be information. honeycomb-lattice to materials, transport shown and is freedom carry of degree valley ehnclengineering, mechanical modes, line oa iiao h lcrncbn tutr nthe sys- in multi-valley , structure traditional as some band such In electronic tems, the to the space. refers of reciprocal which minima freedom, of local degree Bloch the valley manipulating electrons’ by devices electronic dissipation alyrnc eie aebe rpsdi graphene in proposed been nanostructures have devices valleytronics eln lcrncadmcaia properties mechanical and electronic cellent graphene eety uhkn fseilsprel r losonto shown also More are supercells graphene. special doped of kind periodically such in recently, realized be could eety e alyegneigmcaimi pro- is mechanism engineering valley new a Recently, alyrnc i odsg ihecec n low- and high-efficiency design to aim Valleytronics rnmsinsetaadvle rcsigo rpeeand graphene of processing valley and spectra Transmission 26 27 a trce uhatnindet t ex- its to due attention much attracted has √ uigXu, Fuming 18–20 ognrt n oto alyplrzdcur- valley-polarized control and generate to 3 u otebn odn ntesuperlattice, the in folding band the to Due o aoue ihitrvle opig .. in e.g., coupling, inter-valley with nanotubes bon n httelna ad fcro aoue a edt h K the to lead backscatter between can without coupling nanotubes tubes the carbon along of propagate bands electrons i.e., linear armch the corresponding that the n find of of characteristics configurations electronic structural the rich in results superlattice ASnmes 21.d 10.e 32.b 73.63.-b 73.23.-b, 81.05.Ue, 72.10.-d, varyi numbers: or PACS voltage gate a applying by tuned be nanotubes. carbon can armchair current the the of on based transistor valley-field-effect aigthe taking 4 N A e aoaoyo togyCuldQatmMte Phys Matter Quantum Strongly-Coupled of Laboratory Key CAS enmrclyivsiaeteeetoi rnpr prope transport electronic the investigate numerically We 12–16 .INTRODUCTION I. √ × oooia iedefects, line topological K/K 3 √ n yegtcInvto etro unu nomto an Information Quantum of Center Innovation Synergetic and N tlzn,eg,zga edges, zigzag e.g., utilizing, 1 2 3 olg fPyisadEeg,Seze nvriy Shenzh University, Shenzhen Energy, and Physics of College 1–3 eateto hsc n etro hoeia n Computa and Theoretical of Center and Physics of Department nvriyo cec n ehooyo hn,Hfi Anhui Hefei, China, of Technology and Science of University nvriyo cec n ehooyo hn,Hfi Anhui Hefei, China, of Technology and Science of University × N ′ √ 1 aly npitn rpeeare graphene pristine in valleys √ 3 hzo Yu, Zhizhou 24,25 3 , r3 or × CD ee ainlLbrtr o hsclSinea Mic at Science Physical for Laboratory National Hefei ICQD, 3 N √ K rpeespratc sa xml,w hwta tailorin that show we example, an as superlattice graphene 3 N r3 or swl stemperature as well as and ueltie ihitrvle coupling inter-valley with superlattices × K h nvriyo ogKn,Hn og China Kong, Hong Kong, Hong of University The 4 N K 3 and n , and 2 N × ′ . ae Ren, Yafei 9–11 3 K 21–23 ueltie of superlattices N ′ aly fpitn rpeeby graphene pristine of valleys superlattices Particularly, 6–8 tanand strain 17 Various ,4 3, 5 zero- i Wang, Bin the √ e.I,tetgtbnigHmloino h bulk the of Hamiltonian tight-binding the II, Sec. gate scale. large energy at electron current and voltage valley-polarized fully nearly erates aoue r hw nSc I.Abifsmayi given is summary IV. brief A Sec. III. in walled Sec. in single shown and are ribbons, nanotubes armchair nanostructures, and graphene zigzag intro- confined including are transport various and of method electronic on properties function results numerical Green’s The duced. the and graphene pris- of consisting and we tine transistor processing, valley-field-effect valley a employing induce propose By to barrier. coupling of inter-valley presence the the in even nanotubes The superlattices. these in univer- features are general- mechanisms sal of coupling inter-valley loss the without since ity nanotubes carbon and graphene ueia eut hwta hr xs li tunneling- Klein in exist phenomena there like that Our show model. results tight-binding numerical the sin- their using investigate zigzag properties theoretically of and electronic and kinds armchair nanotubes, two walled typical ribbons, gle zigzag and of ribbon, kinds armchair three nanotubes. on and focus nanoribbons We graphene of structural configurations multiple introduces top-adsorption with lattice u td ocro aoue.Wti top-adsorption Within representative the study nanotubes. we carbon case, extend to and study valleytronics our in superlattices graphene these peri rpeepoiiyculdwt topological with substrates. proximity-coupled insulator graphene in appear 3 h eann forppri raie sflos In follows. as organized is paper our of remaining The nti ril,w xlr h oeta plcto of application potential the explore we article, this In N 1 nrbosadnntbs fe studying After nanotubes. and anoribbons n vni h rsneo are.Deto Due barrier. of presence the in even ing i n izgnnrbo emtis we geometries, nanoribbon zigzag and air × aogWei, Yadong aoue hr h alypolarization valley the where nanotube, √ te fgahn aoibn n car- and nanoribbons graphene of rties √ 3 gtelnt fteamhi carbon armchair the of length the ng N 3 × √ n 3 and entneln-iephenomenon, tunnelling-like lein c,adDprmn fPhysics, of Department and ics, √ 3 × rcarcro aoue hc gen- which nanotube, carbon armchair 3 unu Physics, Quantum d 1, √ n586,China. 518060, en N √ uecl,w rps a propose we supercell, 3 ∗ 28 306 China 230026, 306 China 230026, 3 × inlPhysics, tional n hnu Qiao Zhenhua and × 3 roscale, N √ h ukgraphene bulk the g rcaradzga carbon zigzag and armchair 3 ueltie.By superlattices. abnnanotube carbon √ 3 × √ √ ueltieof superlattice 3 ,4, 3, 3 × † √ super- 3 √ 3 × √ 3 2

II. TIGHT-BINDING HAMILTONIAN AND (a)3 (b) THEORETICAL FORMALISM 2 1 1

0 (eV)

f

In √3 √3 graphene superlattice, carbon atoms are -1 2

× E-E still the majority and π-orbital expansion is therefore a -2 3 y reasonable starting point. The nearest-neighbor tight- -3

√ √ -4 binding Hamiltonian of the bulk 3 3 graphene reads:27 × M K M x (c) (d)

† † H = X ti,j c cj + X ui c ci, (1) − i i hi,ji i

† where ci (ci) is the π-orbital creation (annihilation) op- erator on site i. To recover the single-valley metallic phase of bulk √3 √3 graphene superlattice,27 the sys- tem parameters are× precisely determined. The nearest- FIG. 1: Panel (a): band structure of bulk √3 √3 graphene × neighbor hopping amplitudes ti,j are t1 = 2.9 eV be- superlattice calculated from tight-binding Hamiltonian de- tween top-absorption site and adjacent carbon atoms, fined in Eq.1 with periodic boundary conditions. Panel (b), (c) and (d): schematic plots of three types of √3 √3 zigzag and t2 =2.6 eV between carbon atoms, respectively. The × on-site potentials u of different sites are chosen to be graphene ribbons. Green dots stand for top-absorption sites i in the lattice. Blue rectangles indicate the unit cell in every u1 = 4.79 eV for top-absorption site, u2 = 1.35 eV − − setup. The ribbons are finite in y direction and periodically for its three nearest carbon sites, and u3 = 1.05 eV for infinite in x direction. the rest carbon sites with no top-absorption− neighbors. These parameters perfectly recover the band structure in Ref. [27], where periodic adatom or top-absorption III. NUMERICAL RESULTS AND introduces symmetry-breaking and valley-scattering in DISCUSSIONS graphene superlattice. Detailed procedures on param- eter selection can be found in Ref. [27]. In this Section, numerical results on the electronic Based on this tight-binding Hamiltonian, the band structures and transport properties of typical nanorib- structure of bulk √3 √3 graphene is obtained as shown bons and nanotubes of √3 √3 graphene supercell are × ′ in Fig. 1(a) where the inequivalent K and K valleys in presented, including both zigzag× and armchair geome- pristine graphene are folded at the same Γ point due to tries. the band folding. In our transport study, we focus on an energy interval within several eV around the Fermi energy. In Fig.1(a), it is obvious that the three bands A. Zigzag ribbons of √3 √3 graphene superlattice (labeled as ”1”, ”2” and ”3”) dominate in this energy × range. Bands 2 and 3 form an ideal quadratic crossover.27 Moreover, one can also notice that bands ”1” and ”3” As displayed in Fig. 1, we exhibit three kinds of are almost linear around Γ point. The influence of such √3 √3 zigzag graphene nanoribbons (ZGRs). Top- kind of linear dispersion will be discussed when study- absorption× sites are highlighted as green dots. Blue rect- ing the transport properties of √3 graphene nanoribbons angles are used to indicate the unit cells in different and carbon nanotubes. configurations. For simplicity, the setups in panels (b), In below, we utilize the Green’s function method to (c), and (d) of Fig. 1 are respectively denoted as ZGR- investigate the valley-related electronic transport prop- 1, ZGR-2, and ZGR-3. Their geometric differences can erties. From the tight-binding Hamiltonian shown in be easily distinguished from the relative position of top- Eq. (1), the transmission coefficient of electrons at en- absorption sites and their presence/absence at the ribbon ergy E can be expressed as: boundaries. The lower boundaries of ZGR-1 and ZGR-2 are purely consisted of carbon atoms while the absorption r a sites of these two samples are locate on different sublat- T (E) = Tr[ΓLG ΓRG ], (2) tices. Differently, adsorption sites appear at the lower boundary of ZGR-3. As displayed in Fig. 1, these ZGRs where Gr = [E H Σr]−1 is the retarded Green’s func- have finite widths along y direction and are periodic along a − r−† tion, and G = (G ) is the advanced Green’s function. x direction. These zigzag ribbons can be realized by cut- r r r Σ = ΣL +ΣR is the self-energy from the left and right ting a large √3 √3 graphene sheet along proper direc- × 29 leads. ΓL/R is the line width function describing the cou- tions, like the fabrication of graphene nanoribbon or pling between the left/right lead and the central scatter- through the lithography method.30 In our first principles ing region, and can be defined as Γ = i[Σr Σa ]. calculation,31 we found that ZGR-1 exhibits the lowest L/R L/R − L/R 3

1 8

wid= 24 (a)

0

1 6 wid= 52

(a) (b) (c) )

2 f -1 wid=100

(eV) f

4

3 -2 T(E E-E

2

-3

1

0

0 1 8

(b) wid= 24

(d) (e) (f) 2

-1

(eV)

wid= 52 f 6 ) f wid=100

-2 3 E-E

4

-3 T(E

1

2

0 1

0

(g) (h) (i)

(eV) 8 -1 2 f

wid= 24 (c)

wid= 52 -2 3 E-E 6 ) f wid=100

-3

4

- 0 0 0 /- /- T(E k a k a k a

x x x

2

FIG. 2: Band structures of three types of √3 zigzag ribbons 0 with various system sizes. Panels (a)-(c): band structures of -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5

E (eV)

ZGR-1 with ribbon width 24, 52, and 100 lattice sites; Panels f (d)-(f): dispersion relations of ZGR-2 with system size 24, 52, and 100 sites; Panels (g)-(i): bands of ZGR-3 with width 24, FIG. 3: Transmission coefficient as a function of Fermi energy 52, and 100 sites. for three types of √3 √3 zigzag graphene ribbons (ZGRs) at different widths. Panel× (a), (b), and (c) corresponds to ZGR-1, ZGR-2, and ZGR-3, respectively. In all panels, three Gibbs free energy, while ZGR-3 exhibits the highest one. ribbon widths 24, 52, and 100 sites are evaluated. Therefore, ZGR-1 should theoretically be the most stable structure in these ZGRs. In our following discussions, all these setups have been considered and we will show that they possess distinct electronic properties. higher energy degeneracies. Compared with Fig. 1(a), Figure 2 displays the band structures of three different one reasonable explanation is that these envelopes re- zigzag graphene ribbons for different widths (e.g., 24, 52, flect the three bands in the corresponding bulk bands. and 100 lattice sites). The first, second and third rows For ZGR-2, bands for system widths 24, 52, and 100 of Fig. 2 correspond respectively to ZGR-1, ZGR-2 and sites are respectively plotted in Fig. 2(d), (e) and (f), ZGR-3. One can find that Fig. 2(a) (for ZGR-1 with a which exhibit similar characters as those in ZGR-1 ex- width of 24 lattice sites) exhibits a large energy gap ∆ cept that these two edge modes are separable and di- located in the interval of [ 1, 2] eV, deeply underneath vide the band gap into three narrow ones. As the ribbon the Fermi energy. Meanwhile,− − two nearly flat bands lie in width increases, the upper and lower gaps close first and the gap. By projecting the local density of states of the the middle gap disappears at last. The cases for ZGR- two bands onto the lattice sites, we find that they are lo- 3 is similar to that of ZGR-2 as displayed in the last calized along the zigzag ribbon boundaries and thus they three panels of Fig. 2. However, different from ZGR-2, are edge modes, which originate from the dangling bonds the upper surface band of ZGR-3 is more flat. Neverthe- along the zigzag edges, like those in zigzag ribbon of pris- less, the band structures of these three types of zigzag tine graphene. As clearly presented in Fig. 1, the upper ribbons share some common features, including surface bands pinning at k = π and bands envelopes at large and lower edges of these ZGRs are different, leading to x ± the formation of two different edge modes. As the ribbon system sizes. Besides, these zigzag ribbons are all metal- width increases, the energy gap above the edge modes lic since the Fermi level lies in the conduction band. vanishes first as shown in Fig. 2(b) and the lower gap In Fig. 3, we plot the transmission coefficients of these closes at a larger ribbon width as displayed in Fig. 2(c). zigzag ribbons as a function of energy, where the left and Therefore, these two band gaps arise from the finite-size right leads are exactly extended from the central region, effect and disappear when the system size is sufficiently hence resulting in the quantized T (EF). Panels (a), (b) large. and (c) correspond respectively to ZGR-1, ZGR-2 and For ZGR-1 as the system width increases, another ob- ZGR-3. In each setup, three ribbon widths (i.e., 24, 52 servation is that the edge modes are pinned at k = π and 100 lattice sites) are considered. One can see an x ± but evolve with the bulk states around kx 0. One can exact mapping between the transmission coefficients in also notice that, in Fig. 2(c), the band structure≈ estab- Fig. 3 and the band structures in Fig. 2. At small sys- lishes three envelopes, labeled as “1”-“3”. Upon larger tem size, all the ZGRs have zero transmission coefficients system size, the envelopes become further enhanced with at certain energy regions below the Fermi level, where 4

(a) (b) 1 1

y 0 0

(a) (b) (c)

-1 -1

(eV)

x f

wid=24

wid=60 -2 -2 E-E

-3 -3

1 1

0 0

(d) (e) (f)

(eV) -1 -1 f

wid=24

E-E -2 wid=60 -2 FIG. 4: Sketches of two types of √3 armchair graphene rib-

-3 -3

bons: AGR-1 and AGR-2, respectively. Blue rectangles show - /- 0 0 0 2 4 6 8 10 12

the corresponding unit cells and green dots indicate top- k a k a

T(E ) x x absorption sites. The ribbons have finite widths in y direction f and extend to infinite x. ± FIG. 5: Panels (a) and (b): band structures of AGR-1 with system widths 24 and 60 sites. Panel (c) plots the correspond- the dispersion relations show energy gaps. For ZGR-1, ing T vs EF curves for these system sizes. Panels (d) and (e): there are two zero-transmission-coefficient regions while band structures of AGR-2 at ribbon widths 24 and 60 sites. Panel (f) shows the transmission of AGR-2. three gaps exist in ZGR-2 and ZGR-3 at ribbon width (i.e., with 24 lattice sites in black lines). At system size of 52, the zero conducting ranges shrink in all panels of Fig. 3. Only one gap are present in the ZGR-2 and ZGR- 3, and two gaps still reside in the transmission spectrum of ZGR-3, shown in red lines. For a larger system with 100 lattice sites, ZGR-1 has no zero transmission area in lations, two ribbon widths (24 and 60 lattice sites) are the whole energy interval and single narrow gaps appear considered for both AGR-1 and AGR-2. One notices that in panels (b) and (c) of Fig. 3. These finite-size gaps both armchair ribbons are good conductors, similar as eventually disappear at even larger systems. As the rib- the ZGRs. Figures 5(a) and 5(b) plot the band struc- bon width increases, transmission coefficient at the same tures of AGR-1. Apparently, there is a direct band gap at energy increases rapidly for all systems. We also observe kx = 0 around E 1.4 eV. As the system size increases ∼− some oscillation behavior of T (EF) around EF 0.5 eV, from 24 to 60 lattice sites, this gap becomes narrower as which can be attributed to the band overlapping≈ at these shown in Fig. 5(b), indicating its finite-size nature. The energies. One can deduce from Fig .2 that, the oscilla- transmission coefficient of AGR-1 as a function of energy tions tend to be more wild at larger system sizes, which is displayed in Fig. 5(c). It is found that the conducting- is confirmed by our transport calculations. forbidden region matches exactly the band gaps in the left panels. This gap is gradually reduced when system size increases and finally vanishes as the ribbon width B. Armchair ribbons of √3 √3 graphene is large enough. For AGR-2, the numerical results are superlattice × drawn in the lower panels of Fig. 5. The energy disper- sion of AGR-2 is rather similar to that of AGR-1 except that AGR-2 has a larger band gap at the same system Depending on the relative positions of top-absorption size. The transmission coefficient versus energy curves on the honeycomb lattice, there are two different config- in Fig. 5(f) also confirm this observation. As a result, a urations of armchair graphene nanoribbons (AGRs). We wider ribbon width is required to close the band gap in schematically plot these AGRs with widths of two full AGR-2. unit cells in Fig. 4. For simplicity, we denote the armchair ribbon shown in Fig. 4(a) as AGR-1 and the other one as AGR-2. Unlike zigzag ribbons of √3 √3 graphene, both Compared with ZGRs, there is no edge mode in AGRs. × armchair ribbons reproduce perfect √3 √3 periodicity. This is a much natural expectation for armchair-edged × The first-principles calculation suggests that the Gibbs ribbons because of the valley mixture behaviour. An- free energy of AGR-1 is higher than that of AGR-2.31 other striking difference from ZGRs is that there is less Combining with the results of √3 √3 ZGRs, one can oscillation feature in the transmission spectrum of AGRs, × conclude that the ribbons for both zigzag and armchair which can be attributed to the absence of band-folding forms are less stable when the top-absorption sites reside from 1 1 to √3 √3 supercells in the armchair con- on the ribbon boundaries. figurations.× From× Figs. 5(b) and 5(e), one can observe The electronic properties of these AGRs are numeri- that three band envelopes develop to reproduce the bulk cally investigated and displayed in Fig. 5. In our calcu- bands of √3 √3 graphene superlattice. × 5

1 (20, 20) armchair carbon nanotubes. When transforming

0 the zigzag ribbons into armchairs tubes, it is clear that

-1

there are neither edge modes nor gaps in the bands of (eV) f

(a) (b) √ √ -2 3 3 armchair nanotubes. In Fig. 6(a), one can see

E-E ×

that there is a band touching at E 1.4 eV. Near the -3

1 touching point, the band density is≈− low for this (12, 12)

0 armchair tube. For a larger (20, 20) tube as shown in (eV)

-1 f Fig. 6(b), the band density becomes denser with a fixed

(c) (d)

-2 band touching point. The band structures of the two E-E √ √ -3 3 3 zigzag carbon nanotubes are plotted in Fig. 6(c)

0 0 - /- and× 6(d). The band touching also appear in these sys- k a k a

x x 8 8 tems. For the same energy, the bands of zigzag tubes

6 (f) 6 (e) under the touching point reside around kx = 0 point, ) f

4 4 instead of spreading in the whole Brillouin zone as in

T(E (12,12) (12,0) the armchair tubes. This behavior indicates that cur-

2 2

(30,30) (45,0) rent carrier in zigzag ribbons has a larger group velocity

0 0

-3 -2 -1 0 1 -2 -1 0 1 in this energy range. Except this minor difference, the

E (eV) E (eV) bands between armchair and zigzag tubes share lots of f f similarities. FIG. 6: Panels (a) and (b): band structures of √3 √3 Figures 6(e) and 6(f) display the transmission coeffi- armchair carbon nanotubes fabricated from ZGR-3 shown× in cients as a function of energy for the √3 √3 armchair Fig. 1(d), at tube sizes (12, 12) and (20, 20). Panels (c) and and zigzag carbon nanotubes. The results× for (12, 12) (d): dispersion relations of (12, 0) and (45, 0) zigzag carbon and (30, 30) armchair nanotubes are shown in Fig. 6(e). nanotubes, formed by folding armchair ribbons (Fig. 4) in y In a wide energy range of EF [ 2.5, 1) eV, there is direction. Panel (e) plots the T (EF) curves of (12, 12) and only single conducting channel∈ in− the(12− , 12) armchair (30, 30) armchair tubes at different Fermi energies. Panel tube. This energy range corresponds to the two touched (f) shows the transmission of zigzag tubes with system sizes (12, 0) and (45, 0). bands in Fig. 6(a) for the same system parameters. The increase of system size narrows this T = 1 region, as one can see from the red line for the (30, 30) armchair nan- otube. Another interesting finding in Fig. 6(e) is that C. Typical single wall carbon nanotubes of √3 √3 × the transmission coefficients increase by 2 at lot of en- graphene supercell ergy points. This behaviour reveals that many energy levels in the armchair tubes are doubly degenerate, re- The carbon nanotubes of √3 √3 graphene super- gardless of the system size. The transmission spectra of cell have similar structures as pristine× carbon nanotubes. two √3 √3 zigzag tubes with sizes of (12, 0) and (45, 0) × Here we consider two representing configurations: sin- are displayed in Fig. 6(f). Despite their distinct geo- gle wall armchair and zigzag carbon nanotubes. The- metric configurations, the √3 √3 zigzag and armchair × oretically, √3 √3 armchair carbon nanotubes can be carbon nanotubes have similar T (EF) profiles. We also formed by connecting× the upper and lower boundaries of observe the large single mode region for small tube size ZGR-3 as displayed in Fig. 1(d) with proper bonding. and lots of doubly degenerate bands at zigzag nanotubes. Meanwhile, √3 √3 zigzag nanotubes can be formed by In Figs. 6(a) and 6(c), the linear bands parallel × rolling up armchair ribbons as displayed in Fig. 4 and to the blue arrows are rather attractive. In single linking their edges accordingly. Carbon nanotubes con- layer graphene, the massless Dirac fermion with lin- structed in these ways possess the full √3 √3 periodic- ear energy dispersion leads to the counterintuitive Klein × ity. In our consideration, we adopt the classification rule paradox,34,35 where incoming relativistic particles pen- of pristine single wall carbon nanotubes32,33 to denote etrate high barrier with nearly perfect transmission at these two types of √3 √3 carbon nanotubes, where no- certain angles. Klein tunnelling behaviour has been × tations (n,n) and (n, 0) stand for armchair and zigzag revealed in various graphene-based systems, such as tubes, respectively. Integer n in these notations refers graphene p n junctions,36,37 deformed single layer,38 to the number of unit vectors defined in the honeycomb and twisted− bilayer graphene,39 as well as spin-related lattice of pristine bulk graphene. graphene system.40 Similar perfect transmission is also Figure 6 displays the band structures of these √3 √3 observed when electron propagates in pristine carbon armchair and zigzag carbon nanotubes at different× tube nanotubes.41–43 Inspired by these findings, we carry out sizes where one can find that, similar to √3 √3 graphene numerical calculation to show the existence of Klein para- nanoribbons, these two types of carbon× nanotubes are dox in the √3 √3 carbon nanotubes. × also metallic and their Fermi energies lie deeply into the Our system in consideration is a √3 √3 (12, 12) arm- × conduction bands. Specifically, Figs. 6(a) and 6(b) dis- chair nanotube. A potential barrier with height of V0 and play the energy dispersion relations of the (12, 12) and certain length is exerted on the tube while other parts of 6

4 to quantized transmission coefficient T = 1. These nu- merical evidences strongly suggest the existence of Klein

V = 0.0 eV 0 tunnelling-like behavior in the √3 √3 armchair carbon

1 V = 0.1 eV

0 × 3 nanotubes, where electrons transport without backscat-

V = 0.2 eV

0

T(BL) tering in the system, independent of barrier length and

V = 0.5 eV

E = -2.1 eV 0 f heights up to half an electron volt. Similar perfect tun- )

V = 0.5 eV f

0 2 nelling is also observed in the √3 √3 zigzag carbon

0 ×

T(E 0 20 40 60 80 100 nanotubes. However, when the tube size increases, bulk

Barrier length(unit cell) states arise in the whole energy range and gradually co-exist with the linear bands. As a result, the Klein 1 tunnelling-like phenomenon becomes obscure and even- tually disappears.

0

-2.7 -2.4 -2.1 -1.8 -1.5 -1.2 -0.9 -0.6 D. Valley processing in √3 √3 armchair E (eV) × f nanotubes

√ FIG. 7: Transmission vs Fermi energy of a 3 (12, 12) arm- Due to the band folding of the √3 √3 superlattice, chair where a potential barrier with a length the inequivalent K and K′ valleys in× pristine graphene of 10 unit cells exists. Various colored lines stand for different barrier heights V0. Inset: transmission of the same system at and carbon nanotubes are folded into Γ point. There- fixed energy EF = 2.1eV and barrier height V0 = 0.5eV for fore, the intervalley coupling and valley-orbit coupling ef- several barrier lengths.− The length scale ’unit cell’ is indicated fects emerge in these nanotubes of √3 √3 superlattices 27 × in Fig.1(d). of graphene, which qualify them as potential valley- processing materials. Here, we propose a heterostruc- ture composed of 1 1 and √3 √3 armchair carbon × × the system remains unchanged. We calculate the electron nanotubes to act as a valley filter or valley polarizer as transmission coefficient through the barrier. Figure 7 illustrated in Fig. 8(a), which consists of two identical plots the transmission coefficient as a function of energy leads made of pristine armchair nanotubes and a cen- tral scattering region made of √3 √3 armchair carbon for several different barrier heights. The barrier length × is fixed at 10 unit cells, and this length scale of √3 √3 nanotubes. A gate voltage is applied in the central re- armchair tube is depicted in Fig. 1(d), containing 4× sites gion. We consider typical (8, 8) armchair nanotubes at along the x direction. It is found that, in a wide en- both parts, whose energy dispersions are respectively dis- ergy window of E [ 2.5, 1.7] eV, electrons in these played in Figs. 8(b) and 8(c). The intervalley coupling ∈ − − and valley-orbit coupling mechanism in √3 √3 super- armchair carbon nanotubes can almost perfectly pene- × trate the potential barrier. Quantized transmission coef- lattices can manipulate valley polarization coherently in analogy to real spin for . Electrons propagat- ficients of T (EF) = 1 can be achieved for barrier height up to V =0.5 eV. Outside this region, the transmission ing in the left pristine armchair carbon nanotube con- 0 ′ coefficient dramatically drops in the presence of potential tain equivalent K and K valley components. When the barrier. valley-unpolarized current in pristine armchair CNT en- ters the central √3 √3 armchair nanotube region, these We carefully examined the numerical results and found mechanisms break× the balance between the two compo- that, the T (E ) = 1 energy window E [ 2.5, 1.7] eV F nents by flipping electrons of K valley to K’ valley or vice corresponds to the linear dispersion region∈ − in Fig.− 6(a), versa. Hence in the outgoing current to the right lead, suggesting a direct correlation between linear disper- one valley component is larger than the other. In another sion and nearly perfect transmission. In energy range word, the current is valley-polarized. Similar to the spin E [ 2.5, 1.7] eV, the linear dispersion relation guar- polarization, a valley polarization function can be defined antees∈ − the electrons− high group velocity, or so-called rel- to evaluate the efficiency of the device. The valley polar- ativistic electron. In our carbon nanotube system with ization can be adjusted by external factors, such as bias potential barrier, the incident electrons normally collide voltage and gate voltage. The setup presented here can with the barrier. The high-velocity relativistic electron serve as a prototype valley field effect transistor, where can penetrate the barrier without back-scattering, re- valley-polarized current is turned on/off via applying the sulting almost perfect transmission or reflectionless tun- gate voltage as illustrated below. neling. We further checked the dependence of this al- The efficiency of this valley field effect transistor is most perfect tunnelling on barrier length, at fixed elec- characterized by the valley polarization function, which tron energy of E = 2.1 eV and barrier strength of F can be defined in terms of the valley-specified transmis- V = 0.5 eV. The transmission− function versus barrier 0 sion function as: length is plotted in the inset of Fig. 7. Regardless of the barrier length, the system with longer barrier remains TK TK′ PV = − (3) transparent to the incident electrons and also gives rise TK + TK′ 7

V (a) gate 1.0 1.0

0.8 0.8 K K

V (c) (a) (8,8) tube (12,12) tube ... 0.6 0.6 ... Length of 3 tube = 3 units Length of 3 tube = 3 units /P

K'

K’ K’ 0.4 T T 0.4

K' y K' /T K T T

K K

T 0.2 0.2

P P

v x v

0.0 0.0

1.0 1.0

(b) 2 (c) 2

0.8 0.8

1 1 V

0.6 0.6 /P

(b) (8,8) tube (d) (12,12) tube K'

0 0

(eV) 0.4 0.4

f Length of 3 tube= 6 units Length of 3 tube= 6 units /T K T

0.2 0.2 -1 -1 E-E

K K' 0.0 0.0

-2 -2

-0.6 -0.4 -0.2 0.0 0.2 0.4 -0.4 -0.2 0.0 0.2 0.4 0.6

0.0 0.5 1.0 1.5 2.0 -1.0 -0.5 0.0 0.5 1.0

E (eV) E (eV)

k a ( ) k a ( )

x x FIG. 9: Panels (a) and (b): Valley-specified transmission FIG. 8: Panel (a): Illustration of a valley-field-effect- functions and valley polarization versus energy of (8, 8) arm- transistor based on pristine and √3 armchair carbon nan- chair carbon nanotube valley-field-effect-transistor. Lengths otubes. A gate voltage is applied on the √3 armchair carbon of the √3 armchair carbon nanotube are respectively 3 and nanotube region, which functions as valley-processing unit. 6 units. Panels (c) and (d): The same functions for (12, 12) The √3 armchair carbon nanotube is characterized by the armchair carbon nanotube valley-field-effect-transistor. Gate blue-colored top-absorption sites. Panels (b) and (c): band voltage is set to be zero in the calculations. structures of (8, 8) pristine and √3 √3 armchair carbon nanotubes, respectively. × K and K′ valley modes45,46

† ′ ′ ~ ′ ΓL,K/K =ΛK/K ΓL,K/K ΛK K′ , e / where T and T ′ are transmission functions of electrons ′ K K where ΛK,K′ is the eigenfunction of ΓL for K/K valley belonging to K and K′ valleys, respectively. To separate mode. ΓR,K/K′ of the right lead can be produced simi- electrons from equivalent valleys, a simple and effective larly. Following Eq. (2), the valley-specified transmission way is to consider their different group velocities. The 1 ∂E function can be straightforwardly expressed as group velocity, v = ~ ∂k , is related to the band structure, i.e., the slope of the band structures. When focusing on r a TK/K′ = Tr[ΓL,K/K′ G ΓR,K/K′ G ] the first subband of (8, 8) pristine armchair carbon nan- otube shown in Fig. 8(b), it is clear that above the Fermi Then one can calculate the valley polarization PV via energy electron in K valley has larger group velocity than Eq. (3). ′ that of K valley. The situation becomes reverse below We first evaluate the valley-specified transmission co- the Fermi level. Thus, we can calculate the transmission efficient TK/K′ and valley polarization efficiency PV as function contributed from any specific valley under the function of the electron energy. The system under inves- Green’s function frame. In the semi-infinite lead of pris- tigation is armchair carbon nanotube-based valley-field- tine armchair carbon nanotube, the velocity of incident effect-transistor as schematically plotted in Fig. 8(a). electrons is connected with the line width function ΓL The numerically calculated transport properties are ex- ~v U† U 44,45 v in the form of L = ΓL = ΓL . Here L is a hibited in Fig. 9. The energy interval of interest is the diagonal velocity matrix with nonzeroe diagonal elements U first subband of pristine armchair carbon nanotube. For contributed by electrons incoming from the left lead. the (8, 8) armchair system with 3 units length of √3 √3 is a unitary transformation matrix ranked by eigenfunc- × armchair carbon nanotube, both TK and TK′ are continu- tions of ΓL, transforming it into a diagonal matrix ΓL. ous functions of the electron energy as shown in Fig. 9(a). e Obviously, there is an exact mapping between the inci- One can see that TK is rather smooth and larger than dent electron velocities and the eigenvalues of line width 0.8 in the focused energy regime, while the magnitude v ~ function: L = (1/ )ΓL. Considering only the propagat- of TK′ changes more abruptly and exhibits a fluctuating ing modes of the firste subband of (8, 8) armchair carbon pattern. This observation reveals that incident electrons nanotubes [See Fig. 8(b)], both vL and ΓL are 2 2 diag- from K valley of pristine nanotube are less affected by onal matrices. Incident electron from Kevalley has× larger the central √3 √3 armchair carbon nanotube. And velocity, corresponding to the larger one of the two eigen- this fact holds for× all systems considered in Fig. 9. The values of ΓL. Based on this analysis, we can construct calculated valley polarization PV is very small below the effective line width function using only the propagating Fermi level and grows with the increasing electron energy. 8

PV also fluctuates like TK′ in the whole region, and its

P maximum reaches about 0.3 for our considered system. v

0.6 -0.10 When the length of central scattering region increases

0.05 from 3 units to 6 units, TK is still smooth as plotted

0.4 in Fig. 9(b), but TK′ fluctuates more frequently in the 0.15

same energy window. Therefore, the resulting valley po- 0.25 larization vibrates with energy for both below and above 0.2

0.40 the Fermi level. This result suggests that, by increas-

0.45 ing the length of central region, one can realize a more 0.0

0.55

effective manipulation of valley polarization in a rela- E (eV)

tively small energy range. Considering a larger system, -0.2 0.65 such as (12, 12) armchair carbon nanotube valley-field- effect-transistor, an interesting and important observa- 0.75 -0.4 tion is: both TK and TK′ , as well as Pv, are identically 0.85

the same as those in (8, 8) system in the energy win- 0.95 dow of [ 0.6, 0.6] eV, as shown in Fig. 9(c) and 9(d). -0.6 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 We have− examined setups from (6, 6) to (12, 12) carbon

V (eV) nanotube systems and reached the following conclusion: g as long as the electron energy is in the first subband of the pristine armchair carbon nanotube, both the valley- FIG. 11: Valley polarization function PV of a (8, 8) armchair specified transmission functions and valley polarization carbon nanotube valley-field-effect-transistor with 3 units are independent of the circumference of the system. This length, in terms of the electron energy and gate voltage. fantastic property guarantees a great freedom in fabri- cating such kind of valley-field-effect-transistor, since the device’s performance is independent of its transverse di- pristine armchair carbon nanotube. The valley polariza- mension. tion PV as a function of gate voltage Vg is calculated at Secondly, we investigate the influence of the gate volt- two energy points and different system sizes as shown in age on the device performance. The gate voltage is ap- Fig. 10. From Fig. 10(a), one can find that the valley po- plied on the √3 √3 armchair carbon nanotube region, larization fluctuatingly grows as the applied gate voltage which serves as the× valley-processing unit. In our calcu- decreases from positive to negative. For a (8, 8) valley- field-effect-transistor with 3-unit length of the √3 √3 lation, gate voltage simply shifts the on-site energies of × the affected atoms, i.e., diagonal elements of their Hamil- armchair carbon nanotube, the valley polarization can tonian. Electron energy is kept at the first subband of reach about 0.9 at the negative gate voltage, which in- dicates that the K valley is almost fully polarized. The wide PV > 0.9 plateau shown in Fig. 10(a) qualifies the device as a stable valley-polarized current generator that 1.0 1.0 v (8,8) (8,8) can operate in a broad gate voltage range. L=3 E=0.36 eV L=6 E=0.25 eV

0.8 0.8

(a) (b) (8,8) (8,8) More importantly, PV jumps abruptly from below 0.02

L=3 E=0.60 eV L=6 E=0.45 eV 0.6 0.6 to above 0.9 at Vg 0.5 eV, showing a great on/off ≈ −

0.4 0.4 ratio of this valley-field-effect-transistor. The fluctuation of PV exists at all electron energies. Comparing the re-

0.2 0.2

Valley Polarization P sults for E = 0.36 eV and E = 0.60 eV, obviously high

0.0 0.0

(12,12) (12,12) valley polarization can be easily achieved when the sys- v

L=3 E=0.36 eV L=6 E=0.25 eV 0.8 0.8 tem works at larger electron energy, which only requires

(c) (d) (12,12) (12,12) V 0.25 eV to reach P > 0.9 at E =0.60 eV. When L=3 E=0.60 eV L=6 E=0.45 eV g V 0.6 0.6 increasing≈− the length of the central regime to 6 units

0.4 0.4 in the device, more fluctuations of PV are revealed in

0.2 0.2 Fig. 10(b), suggesting a more effective gate modulation. But the high valley polarization plateau shrinks at all Valley Polarization P

0.0 0.0

-0.9 -0.6 -0.3 0.0 0.3 0.6 0.9 -0.9 -0.6 -0.3 0.0 0.3 0.6 0.9 electron energies and a larger negative Vg is required for

V (eV) V (eV) PV > 0.9. We also performed calculations on a (12, 12) g g armchair carbon nanotube system and the corresponding FIG. 10: Panels (a) and (b): Valley polarization versus gate results are displayed in Fig. 10(c) and 10(d). It is found voltage of (8, 8) armchair carbon nanotube valley-field-effect- again that the device performance is independent of the transistor at several electron energies. ’L’ in the legends diameter of the nanotubes. stands for the lengths of the √3 armchair carbon nanotube, We summarize this part with the working map of a which are respectively 3 and 6 units in the calculation. Panels (8, 8) armchair carbon nanotube based valley-field-effect- (c) and (d): The same functions for (12, 12) tube systems. transistor. Its valley polarization PV as functions of elec- 9 tron energy and gate voltage are displayed in Fig. 11. transmission through the tube is quantized even in the PV function of this device can be effectively tuned by ap- presence of a potential barrier, regardless of the barrier plying an external gate voltage. The best working zone length and height up to a few hundreds of meV. A valley- of this device is the large red triangular area highlighted field-effect-transistor consisting of pristine and √3 √3 in Fig. 11, where the K valley is nearly fully polarized. armchair carbon nanotubes is proposed, which can× be The valley-field-effect-transistor has a stable output in used to filter fully valley-polarized current and can be this zone, and the on/off ratio is guaranteed by the nar- tuned by applying an external gate voltage or adjust- row blue region adjacent to the red zone, whose valley ing the length of the central scattering regime. As long polarization is below 0.05. The larger the electron en- as the electron energy is within the range of the first ergy, the easier to get fully valley-polarized current. The subband of pristine armchair carbon nanotubes, perfor- performance of this device is the same for systems with mance of this valley-field-effect-transistor is independent different circumferences, as long as the energy is within of the tube circumference. the first subband of pristine armchair carbon nanotube.

IV. CONCLUSION V. ACKNOWLEDGMENTS

In conclusion, we have numerically investigated the This work was financially supported by the NNSFC electronic properties of typical √3 √3 graphene (Grants No. 11504240, No. 11574217, No. 11304205, and nanoribbon and nanotube structures. We× show that all No. 11474265), NSF of SZU (Grant No. 201550). Y.R. the √3 √3 nanostructures are metallic materials. Both and Z.Q. also acknowledge the financial supports from the zigzag× and armchair ribbons have finite-size energy the China Government Youth 1000-Plan Talent Program, gap below the Fermi energy more than 1 eV. Double- Fundamental Research Funds for the Central Universities degeneracy in energy levels instead of energy gaps are (WK3510000001 and WK2030020027) and the National found in the spectra of both armchair and zigzag car- Key R & D Program (Grant No. 2016YFA0301700) . bon nanotubes. In small √3 √3 carbon nanotubes, The Supercomputing Center of USTC is gratefully ac- there is a large energy range showing× linear dispersion, knowledged for the high-performance computing assis- which leads to the Klein tunneling-like behavior: electron tance.

∗ Correspondence author: [email protected] (2007). † Correspondence author: [email protected] 15 W. Yao, D. Xiao, and Q. Niu, Phys. Rev. B 77, 235406 1 K. Takashina, Y. Ono, A. Fujiwara, Y. Takahashi, and Y. (2008). Hirayama, Phys. Rev. Lett. 96, 236801 (2006). 16 A. Hill, A. Sinner and K. Ziegler, New Journal of Physics 2 L. M. McGuire, Mark Friesen, K. A. Slinker, S. N. Cop- 13, 035023 (2011). persmith and M. A. Eriksson, New Journal of Physics 12, 17 A. Rycerz, J. Tworzydo, and C. W. J. Beenakker, Nature 033039 (2010). Phys. 3, 172 (2007). 3 D. Culcer, A. L. Saraiva, B. Koiller, X. Hu, and S. Das- 18 Z. Qiao, J. Jung, Q. Niu, and A. H. MacDonald, Nano Sarma, Phys. Rev. Lett. 108, 126804 (2012). Letters, 11, 3453 (2011). 4 Z. Zhu, A. Collaudin, B. Fauqu´e, W. Kang, and K. Behnia, 19 Z. Qiao, J. Jung, C. Lin, A. H. MacDonald, and Q. Niu, Nature Phys. 8, 89 (2011). Phys. Rev. Let. 112, 206601 (2014). 5 J. Isberg, M. Gabrysch, J. Hammersberg, S. Majdi, K. K. 20 Y. Ren, Z. Qiao, and Q. Niu, Rep. Prog. Phys. 79, 066501 Kovi, and D. J. Twitchen, Nature Mater. 12, 760 (2013). (2016). 6 I. Zuti´c,ˇ J. Fabian, and S. Das Sarma, Rev. Mod. Phys. 21 D. Gunlycke and C. T. White, Phys. Rev. Lett. 106, 76, 323 (2004). 136806 (2011). 7 K. Wakabayashi, Y. Takane, M. Yamamoto and M. Sigrist, 22 Y. Liu, J. Song, Y. Li, Y. Liu, and Q.-F. Sun, Phys. Rev. New Journal of Physics 11, 095016 (2009). B 87, 195445 (2013). 8 K. S. Novoselov, Rev. Mod. Phys. 83, 837 (2011). 23 J.-H. Chen, G. Autes, N. Alem, F. Gargiulo, A. Gautam, 9 A. K. Geim and K. S. Novoselov, Nature Mater. 6, 183 M. Linck, C. Kisielowski, O. V. Yazyev, S. G. Louie, and (2007). A. Zettl, Phys. Rev. B 89, 121407(R) (2014). 10 A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. 24 T. Fujita, M. B. A. Jalil, and S. G. Tan, Appl. Phys. Lett. Novoselov and A. K. Geim, Rev. Mod. Phys. 81, 109 97, 043508 (2010); Z. Khatibi, H. Rostami, and R. Asgari, (2009). Phys. Rev. B 88, 195426 (2013). 11 S. Das Sarma, S. Adam, E. H. Hwang, E. Rossi, Rev. Mod. 25 F. Zhai, Y. Ma and K. Chang, New Journal of Physics Phys. 83, 407 (2011). 13, 083029 (2011); Y. Jiang, T. Low, K. Chang, M. I. 12 A. F. Morpurgo and F. Guinea, Phys. Rev. Lett. 97, Katsnelson, and F. Guinea, Phys. Rev. Lett. 110, 046601 196804 (2006). (2013). 13 E. McCann, Phys. Rev. Lett. 97, 146805 (2006). 26 X. Chen, L. Zhang, and H. Guo, Phys. Rev. B 92, 155427 14 D. Xiao, W. Yao, and Q. Niu, Phys. Rev. Lett. 99, 236809 (2015); Z. Yu, F. Xu, J. Wang, Carbon 99, 451 (2016). 10

27 Y. Ren, X. Deng, Z. Qiao, C. Li, J. Jung, C. Zeng, Z. (2010). Zhang, and Q. Niu, Phys. Rev. B 91, 245415 (2015). 39 W.-Y. He, Z.-D. Chu, and L. He, Phys. Rev. Lett. 111, 28 W. Cao et al., arXiv:1602.08822. 066803 (2013). 29 L. C. Campos, V. R. Manfrinato, J. D. Sanchez-Yamagishi, 40 M.-H. Liu, J. Bundesmann, and K. Richter, Phys. Rev. B J. Kong, and P. Jarillo-Herrero, Nano Lett. 9, 2600 (2009). 85, 085406 (2012). 30 G. Xie, Z. Shi, R. Yang, D. Liu, W. Yang, M. Cheng, D. 41 C. T. White and T. N. Todorov, Nature 411, 649 (2001); Wang, D. Shi, and G. Zhang, Nano Lett. 12, 4642 (2012). W. Liang, M. Bockrath, D. Bozovic, J. H. Hafner, M. Tin- 31 Z. Yu, F. Xu, Z. Qiao, and J. Wang, unpublished. kham, and H. Park, Nature 411, 665 (2001). 32 T. Ando, J. Phys. Soc. Jpn. 74, 777 (2005). 42 G. A. Steele, G. Gotz, and L. P. Kouwenhoven, Nature 33 J.-C. Charlier, X. Blase, and S. Roche, Rev. Mod. Phys. Nanotech. 4, 363 (2009). 79, 677 (2007). 43 V. Jakubsky, L.-M. Nieto, and M. S. Plyushchay, Phys. 34 M. I. Katsnelson, K. S. Novoselov and A. K. Geim, Nature Rev. D 83, 047702 (2011). Physics 2, 620 (2006). 44 P. A. Khomyakov, G. Brocks, V. Karpan, M. Zwierzycki, 35 C. W. J. Beenakker, Rev. Mod. Phys. 80, 1337 (2008). and P. J. Kelly, Phys. Rev. B 72, 035450 (2005). 36 A. F. Young and P. Kim, Nature Physics 5, 222 (2009). 45 Y. X. Xing, L. Zhang, and J. Wang, Phys. Rev. B 84, 37 N. Stander, B. Huard, and D. Goldhaber-Gordon, Phys. 035110 (2011). Rev. Lett. 102, 026807 (2009). 46 J. Wang and H. Guo, Phys. Rev. B 79, 045119 (2009). 38 O. Bahat-Treidel, Or Peleg, M. Grobman, N. Shapira, M. Segev, and T. Pereg-Barnea, Phys. Rev. Lett. 104, 063901