Probing Spin Helical Surface States in Topological HgTe Nanowires
Johannes Ziegler, Raphael Kozlovsky, Cosimo Gorini, Ming-Hao Liu, Sabine Weishäupl, Hubert Maier, Ralf Fischer, Dmitriy A. Kozlov, Ze Don Kvon, Nikolay N. Mikhailov, Sergey A. Dvoretsky, Klaus Richter, Dieter Weiss
PProbing Spin Helical Surface States in Topological HgTe Nanowires
J. Ziegler, R. Kozlovsky, C. Gorini, M.-H. Liu ( 劉 明 豪 ),
2, 3
S. Weishäupl, H. Maier, R. Fischer, D. A. Kozlov,
4, 5
Z. D. Kvon,
4, 5
N. Mikhailov, S. A. Dvoretsky, K. Richter, and D. Weiss Institut für Experimentelle und Angewandte Physik, Universität Regensburg, 93053 Regensburg, Germany Institut für Theoretische Physik, Universität Regensburg, 93053 Regensburg, Germany Department of Physics, National Cheng Kung University, Tainan 70101, Taiwan A.V. Rzhanov Institute for Semiconductor Physics, Novosibirsk, Russia Novosibirsk State University, Russia (Dated: December 12, 2017)Nanowires with helical surface states represent key prerequisites for observing and exploiting phase-coherenttopological conductance phenomena, such as spin-momentum locked quantum transport or topological super-conductivity. We demonstrate in a joint experimental and theoretical study that gated nanowires fabricated fromhigh-mobility strained HgTe, known as a bulk topological insulator, indeed preserve the topological nature ofthe surface states, that moreover extend phase-coherently across the entire wire geometry. The phase-coherencelengths are enhanced up to 5 µ m when tuning the wires into the bulk gap, so as to single out topological trans-port. The nanowires exhibit distinct conductance oscillations, both as a function of the flux due to an axialmagnetic field, and of a gate voltage. The observed h/e -periodic Aharonov-Bohm-type modulations indicatesurface-mediated quasi-ballistic transport. Furthermore, an in-depth analysis of the scaling of the observedgate-dependent conductance oscillations reveals the topological nature of these surface states. To this end wecombined numerical tight-binding calculations of the quantum magneto-conductance with simulations of theelectrostatics, accounting for the gate-induced inhomogenous charge carrier densities around the wires. Wefind that helical transport prevails even for strongly inhomogeneous gating and is governed by flux-sensitivehigh-angular momentum surface states that extend around the entire wire circumference. I. INTRODUCTION
Three-dimensional topological insulators (3DTIs) are aparticular class of bulk insulators hosting time reversalsymmetry-protected metallic surface states. The latter are he-lical, i.e. characterized by (pseudo)spin-momentum locking,and described by low-energy effective Dirac-type models . Innanowires based on 3DTI materials such locking heavily af-fects the one-dimensional (1D) subband spectrum and, if com-bined with superconductivity, is a basic ingredient for the real-ization of Majorana modes . Moreover, from a general quan-tum transport perspective, 3DTI nanowires provide a particu-larly rich playground due to the interplay between topologi-cal properties and effects arising from phase coherence. Thefact that the conducting states are “wrapped” around an in-sulating bulk, in conjunction with their helical nature, leadsto various interesting and geometry-sensitive magnetoresis-tive phenomena that are inaccessible in standard metallicsystems, whose bulk and surface contributions cannot in gen-eral be singled out.In particular, a 3DTI nanowire in a coaxial magnetic fieldwith magnitude B and associated flux φ = AB , as sketched inFig. 1(a), is expected to show peculiar Aharonov-Bohm typemagnetoresistance features. Indeed, oscillations with a periodof one flux quantum φ = h/e (where h = 2 π (cid:126) is Planck’sconstant and e the elementary charge) were observed in earlyexperiments . According to theory these oscillations re-flect the wire’s 1D subband structure, given by E = ± (cid:126) v F (cid:113) k z + k l with k l = 2 πP (cid:18) l + 12 − φφ (cid:19) . (1)In Eq. (1), v F is the Fermi velocity of the surface carri- ers, k z the coaxial and k l the transversal wave vector, thelatter having the meaning of angular momentum. The angu-lar momentum quantum number is labeled by l ∈ Z and itshalf-integer shift in k l is caused by a curvature induced Berryphase. The resulting energy spectrum is sketched in Fig. 1(b)for three characteristic values of the magnetic flux. For φ = 0 an energy gap is present (due to the Berry phase) and the 1Dsubbands are twofold degenerate with respect to angular mo-mentum. Note, however, that owing to their Dirac-like naturethese states are not spin-degenerate. For finite flux such as φ/φ = 0 . the degeneracies with respect to k l are lifted.For φ/φ = 0 . , the magnetic flux cancels the Berry phaseand the l = 0 states become gapless, k l =0 = 0 . This lin-ear gapless state is non-degenerate. More generally, the totalnumber of states is odd, and time-reversal symmetry, restoredat φ/φ = 0 . , implies one “perfectly transmitted mode” .For φ/φ = 1 the spectrum recovers its φ/φ = 0 form. Fur-ther increasing the flux leads to analogous cycles: the Berryphase cancellation, and thus the appearance of a gapless state,takes place for half-integer values of φ/φ , whenever k l = 0 for a given l , while for integer values the gapped spectrum for φ = 0 in Fig. 1(b) is recovered.This “shifting” of the 1D subbands with changing flux im-plies that for a given Fermi level position the conductanceshould be φ -periodic. Dominating h/e oscillations were in-deed observed in a number of experiments . Furthermore,the phase of the Aharonov-Bohm-type oscillations depends onthe Fermi level position. To be definite, consider E F close tozero: no states are present for φ = 0 (and corresponding evenmultiples of φ / ), whereas for φ/φ = 0 . (and correspond-ing odd multiples of φ / ) the gapless mode emerges. Hence,one expects a conductance minimum (maximum) for integer(half-integer) flux quanta. For E F slightly above zero, e. g. at a r X i v : . [ c ond - m a t . m e s - h a ll ] D ec CdTe
HgTe
CdTe (4 µm) (on (013) GaAs)
HgCdTeHgCdTeAl O AuSiO ϕ / ϕ = 0.25 ϕ / ϕ = 0.5 ϕ / ϕ = 0 k z k z k z E EE
As B , k z (c) (b)(a) (d) Figure 1. HgTe-based nanowire. (a) Sketch of a nanowire with crosssection A and magnetic field (cid:126)B along the wire axis. (b) 1D bandstructures of the surface states of three-dimensional topological in-sulator nanowires for three different representative magnetic fluxes.(c) Schematic cross section of a nanowire employed in experiment.The nanowires are etched from wafers containing an 80 nm strainedHgTe film and consist of a 4 µ m CdTe base layer grown on (013)GaAs substrate, two 20 nm Hg . Cd . Te buffer layers on top andbottom of the HgTe and a 40 nm CdTe cap layer. After wet chemicaletching the wire is covered by SiO /Al O and a metallic top gate.(d) SEM micrograph of a representative HgTe TI nanowire, taken ata tilt angle of 50° before deposition of the topgate structure. It has alength of 1.3 µ m and a median width of 163 nm. the dashed line in Fig. 1(b), the situation is reversed: twomodes are present for φ = 0 , and only one for φ/φ = 0 . .Such π -phase shifts as a function of the Fermi level position, i. e. the gate voltage V g , were observed in Bi . Sb . Se and Bi Te . If the Fermi level can be tuned to the Diracpoint a conductance minimum for φ/φ = 0 and a maximumat φ/φ = 0 . , as observed in , are signatures of the Berryphase of π and thus of spin-helical Dirac states. However, instrained HgTe nanowires investigated here, the Dirac point isburied in the valence band and thus cannot be singled-outand probed on its own. Without direct access to the latter, thephase switching alone is not an exclusive signature of Diracstates: the 1D subband spectrum of trivial surface states, andthus the resulting conductance, would also be φ -periodic.Hence the crucial question arises how to distinguish topo-logical from trivial states in 3DTI nanowires with an inacces-sible Dirac point, such as strained HgTe. This is the centralpoint that we address below, by quantitatively analyzing theconductance oscillation periodicity occurring as a function ofgate voltage V g at fixed flux. The observed oscillations reflectdirectly the 1D subband structure and its degeneracies. Thisallows us to draw conclusions about the nature of the surfacestates, trivial ones being spin-degenerate, in contrast to spin-helical Dirac states. In doing so, we also address a secondissue, namely the consequences of a varying carrier densityaround the wire circumference. Experiments typically rely onthe use of top and/or back gates, which couple differently to device w (nm) l ( µ m) P (nm) ∆ B h/e (T)t1 302 0.99 724 0.203t2 518 2.06 1156 0.116t3 246 2.51 613 0.249w1 310 1.33 740 0.197w2 163 1.33 446 0.386w3 178 1.06 476 0.351w4 294 1.95 708 0.208w5 287 2.97 694 0.212Table I. Geometrical parameters of the nanowire structures consid-ered. The height h of all devices is given by the 80 nm thick HgTelayer of the wafers. Median width and length of the structures aredenoted by w and l , respectively. The circumference P and the ex-pected period for h/e -oscillations ∆ B h/e (in Tesla) were calculatedassuming the surface state wave functions to be located ∼ nm be-neath the bulk surface. the top, bottom and side surfaces of the 3DTI wires. The car-rier density (or the capacitance) therefore becomes a stronglyvarying function of the circumference coordinate s . Note alsothat a certain degree of inhomogeneity is expected even inthe absence of gating, as a consequence of intrinsic systemanisotropies . As we show below, an inhomogeneous surfacecharge distribution modifies significantly the band structure,yet leaves the essential physics intact. II. NANOWIRE DEVICES
The investigated TI nanowires were fabricated from (013)oriented, strained nm thick HgTe thin films grown bymolecular beam epitaxy on (013) oriented GaAs substrates(for details see Ref. ). A cross section through the layer se-quence is shown in Fig. 1(c). Using electron beam lithogra-phy and wet chemical etching, nanowires as the one shownin Fig. 1(d) were fabricated. For etching we used a Br -basedwet etch process to preserve the high charge carrier mobili-ties of the bulk material. The wet etching did not result inperfectly rectangular wire cross sections but rather in trape-zoidal ones with the narrower side on top. In Fig. 1(d) thetop width of the wire’s central segment was nm whilethe bottom width was nm, resulting in an average (me-dian) width of nm. For easier modeling below we usea rectangular cross section with area A given by the averagewire width times HgTe film thickness, A = wh . This ap-proximation reduces the circumference P that determines thespacing between angular momenta, ∆ k l = 2 π/P , by − .As discussed below, the resulting effect of this assumption isnegligible. After etching the wires were covered with nmof Si O using plasma enhanced chemical vapor depositionand nm Al O deposited with atomic layer deposition.For gating a metallic top layer consisting of titanium and goldwas used. The resulting schematic cross section is depictedin Fig. 1(c) and will be used for the electrostatic modelingin Sec. IV. Ohmic contacts to the wire were formed via sol-dered indium. The nanowires were fabricated in both a true4-terminal geometry (devices denoted by w3-w5) as well as a -2 -1 0 1 215.615.816.016.216.416.6 0.0 0.5 1.0 1.502468 (a) G ( e / h ) B (T)
40 mK 100 mK 200 mK h/e G ( e / h ) T (mK) (b) t1 t2 t3 l ( (cid:181) m ) V g (V) T = 40 mK Figure 2. (a) Magnetoconductance oscillations of device t1 mea-sured at different temperatures T . The traces show h/e -periodic os-cillations with amplitudes ∆ G ( T ) decreasing with increasing tem-perature T . The T -dependence of the amplitude of sample t1, takenat various gate voltages, is shown in the inset of Panel (b), alongwith the corresponding fits using Eq. (2). The exponential fits allowfor the extraction of l ϕ , displayed in Panel (b) as a function of gatevoltage V g for devices t1-t3. quasi-2-terminal geometry (devices t1-3, w1-2). In the lattercase, the nanowire constriction was embedded in a larger Hallbar, where the voltage probes are several µ m removed fromthe device.The magnetoconductance of the wires was measured in adilution refrigerator at temperatures of typically 50 mK andmagnetic fields up to 5 T. A rotating sample holder was usedfor measurements to allow for in- and out-of-plane alignmentof the magnetic field. Standard AC lock-in techniques andFemto voltage preamplifiers were used at excitation ampli-tudes and frequencies of typically – nA and – Hz,respectively. Additionally, a cold RC-filter was added to sup-press noise at the top gate and to improve reproducibility ofmagnetotransport traces. In the following, we present resultsfrom a total of 8 nanowires. Parameters defining their geome-tries are listed in Table I. Devices t1-t3 were investigated re-garding the temperature dependence of the Aharonov-Bohm-type oscillations (see Sec. III A), while devices w1-w5 werestudied with regard to signatures of the subband structure (seeSec. III B). -0,5 0,0 0,5 1,0 1,5 2,0 2,5 3,0 R w ( B ) / R w ( ) V g (V) CNP E V E C T=1.5 K
Figure 3. Normalized longitudinal resistance R w ( B ) /R w ( B =0) as a function of gate voltage for nanowire device w1. The upperand lower bounds of the bulk band gap are indicated by arrows andwhere determined by comparing to analysis done on material withthe same wafer stack by Kozlov et al . . III. CHARACTERIZATION OF CONDUCTANCEOSCILLATIONSA. Magnetoconductance
Figure 2(a) shows the measured two-point conductance G as a function of a magnetic field B applied along the wire axis.The experimental data was taken at temperatures T between40 mK and mK. An overall h/e periodicity is clearly vis-ible, as indicated by the horizontal bar in Fig. 2(a) (and an-alyzed in more detail below). The fact that G ( B ) exhibitsAharonov-Bohm-type h/e periodicity instead of h/ (2 e ) be-havior, arising from interference between time-reversed pathsin the diffusive limit , implies that transport along the wiresis indeed non-diffusive, i.e. the elastic mean free path is pre-sumably not much shorter than the wire length and larger thanthe wire circumference. However, the additional conductancefluctuations present in G ( B ) in Fig. 2(a) indicate residual dis-order scattering.From the temperature dependence we can estimate thephase coherence length l ϕ by using the exponential decay ofthe amplitude ∆ G of the h/e conductance oscillations for bal-listic transport on scales of the perimeter P , ∆ G ∝ exp (cid:18) − Pl ϕ ( T ) (cid:19) . (2)In Fig. 2(b) the resulting phase coherence lengths obtainedfrom three devices are plotted as a function of V g . For all sam-ples minimal values of µ m to µ m are found for l ϕ , whilemaximal values reach µ m at gate voltages around V g = 1 V.At V g = 1 V the Fermi level E F is in the bulk gap, as ex-tracted from independent measurements for macroscopic Hallbars made from the the same material . For E F in the gapthe phase coherence lengths of the topological surface statesare expected to be largest as backscattering is reduced andscattering into bulk states is suppressed.Figure 3 shows V g -dependent resistance traces of devicew1 for a number of magnetic fields B normalized to theirvalue at B = 0 . The maximum of the longitudinal resis-tance, which comes out more clearly with higher magneticfields, is usually ascribed to the charge neutrality point (CNP).In strained HgTe it is located slightly in the valence band .Hence the valence band edge is located on the right hand sideof the maximum. The V g locations of the conduction and va-lence band edges, indicated by arrows in Fig. 3, have beenobtained by comparing to the results described by Kozlov etal. (2014) .Next we study in more detail the Aharonov-Bohm-typeoscillations and in particular their V g -dependence, here rep-resentatively discussed for device w1. To this end, we re-move the small hysteresis of the superconducting magnet fromthe conductance and subtract a smoothly varying background G sm ( B ) using a Savitzky-Golay filter . The filter was ap-plied in a way that frequencies smaller than / φ are cut off.The resulting conductance ∆ G ( V g ; φ ) is shown in the colorplot in Fig. 4(a). Here, the gate voltage spacing of neighbor-ing traces is ∆ V g = 0 . V. Corresponding data, but now asline cuts taken at different values of V g are shown in Fig. 4(b)for sample w2. ∆ G shows dominant φ -periodic oscillationsover a large gate voltage range, as well as a pronounced phaseswitching between minima and maxima, i.e. additional con-ductance oscillations upon varying V g at fixed flux (to be ana-lyzed in Sec. III B). The φ -periodicity is confirmed by a fastFourier transform (FFT) analysis of traces covering a mag-netic field range corresponding to 20 φ in the case of devicew1. As the period of the oscillations should be independent ofthe gate voltage an average of the FFTs was taken in the volt-age range to V. The V g -averaged FFT is shown in Fig. 5as a function of /φ . Here, we used the square geometry witha median width of nm for w1 and assumed that the topo-logical surface states are nm below the surface to computethe magnetic flux φ = AB from B . The resulting FFT peakis located at . /φ . The peak falls into the expected regionof an h/e signal, where the lower bound is given by the ge-ometrical dimensions of the nanowire cross section as listedin Table I. The upper bound is calculated for the case wherethe wave function of the surface states lies 8 nm within the TIbulk.To conclude, the distinct peak of the FFT close to /φ implies that, at low temperatures, transport is mediated bystates extending phase-coherently across the entire surface ofthe weakly disordered nanowire. B. Subband-induced conductance modulation
The detection of the Aharonov-Bohm oscillations thatswitch their phase as a function of V g is by itself insufficientto confirm the topological nature of the underlying surfacestates. Therefore we have to go beyond previous analysis andaim at a quantitative description of the ∆ G ( V g ) -oscillations,which directly reflect the subband structure of the quasi-1Dwires. ∆ G ( V g ; φ ) is extracted from line cuts taken at half in-teger and integer values of φ/φ in Fig. 4(a) and plotted as blue, respectively orange, curves in Fig. 6(a). To suppressthe influence of aperiodic conductance fluctuations we aver-aged over all positive and negative integer multiples of φ inthe dataset, i.e. ± φ , ± φ and ± φ to generate the orangecurve in the case of the representative device w1. Accord-ingly, the blue curve results from an average over positiveand negative half-integer multiples calculated from ± . φ , ± . φ and ± . φ . The line cut for φ = 0 was omit-ted due a large weak antilocalization-like background profilethat renders background removal difficult. In Fig. 6(a) thetwo conductance curves ∆ G ( V g ) obtained in this manner ex-hibit antiphase oscillations as a function of V g , i.e. maxima of ∆ G ( V g ) for integer multiples of φ go along with minima of ∆ G ( V g ) for half integer multiples.In Secs. IV-VI we carry out an in-depth analysis associat-ing these characteristics with the quantized subbands of he-lical surface states. Based on this study the analysis of theconductance oscillations in Fig. 6(a) proceeds as follows: Theminima in ∆ G ( V g ) correspond to a Fermi level at the edgeof a subband since the corresponding high density of states(the van-Hove singularity) causes enhanced scattering. Ac-cordingly, the gate voltage distance ∆ V g between two min-ima, corresponding to the period of the conductance oscilla-tion, can be directly mapped onto the subband spacing. Onekey property that distinguishes topological surface states fromtheir trivial counterparts is their degree of spin-degeneracy. Itdirectly affects ∆ V g since a spin-degeneracy of two, as fortrivial surface states, implies that twice as many states needbe filled compared to helical edge states. This allows us toextract the spin degeneracy of the subbands involved by quan-titatively analyzing the distance ∆ V g between adjacent con-ductance minima.To increase the accuracy of the extracted ∆ V g , we usethe anticorrelation between the two curves for jφ and ( j +0 . φ with j = ± , ± , ... and the corresponding minima-maxima pairs in order to label the pairs by a running indexand to obtain ∆ V g . Pairs with the maximum correspond-ing to integer/half-integer flux quantum are labeled by aninteger/half-integer index by black and green dashed lines, re-spectively. Since in strained HgTe the Dirac point is locatedin the valence band, we do not know the precise number offilled 1D subbands at a given gate voltage. We thus startcounting by fixing a reference voltage V at which a certain(unknown) number N of subbands is filled, and use the rel-ative index N − N , with N the total number of occupiedsubbands. In Fig. 6(b) the gate voltage at which a particularminimum-maximum pair occurs is plotted as a function of thecorresponding index N − N .The connection between ∆ G ( V g ; φ ) and the nanowirebandstructure is a priori not obvious, since the nanowire sur-face charge carrier density, and ergo its capacitance, is inho-mogeneous due to asymmetric gating. However, as we willdetail in Secs. IV-VI, an analysis in terms of a single effec-tive charge carrier density n eff , i. e. effective capacitance C eff ,turns out to be fully justified. They are related to each othervia ( n eff − n ) e = C eff ( V g − V ) , (3) V g ( V ) -3 -2 0 2 3 -0.2-0.100.10.20.3-1 1 (b)(a) Figure 4. Quantum conductance corrections of transport measurements in TI nanowires. (a) Color plot of conductance correction ∆ G as afunction of magnetic flux φ (in units of flux quantum φ ) and gate voltage V g from device w1. (b) Selection of representative magnetoconduc-tance curves from device w2 comprising a larger flux range. The different traces (offset for clarity) exhibit clear φ -periodic behavior, as wellas a switching of the phase with varying V g . FFT m agn i t ude ( a . u . )
1/ (1/ ) Figure 5. Fourier spectrum of nanowire magnetoconductance oscil-lations. Fast Fourier transform, calculated from the magnetoconduc-tance of device w1 and averaged over all V g values, shows a domi-nant peak close to /φ , reflecting φ = h/e -periodic oscillations.The bar at the peak denotes the expected range of the peak to occur,see text. where V is the gate voltage at which the oscillation count-ing starts and n the corresponding carrier density. For a 2Dsystem the Fermi wave vector k F is given by k F = (cid:112) (4 π/g s ) n eff (4)with the spin degeneracy factor g s = 1 and g s = 2 for spin-resolved topological surface states and spin-degenerate triv-ial surface states, respectively. Assuming a constant subbandsplitting E/ ( (cid:126) v F ) = ∆ k l = 2 π/P (in view of Eq. (1)), the Fermi wave vector at a conductance minimum (subband open-ing) can be written as k F = k + ( N − N )∆ k l with k = (cid:112) (4 π/g s ) n . (5)Combining Eqs. (3), (4) and (5) yields the relation V g − V = g s e πC eff (cid:2) k ( N − N )∆ k l + ( N − N ) ∆ k l (cid:3) (6)between gate voltage and subband level index N . The dis-tance between two conductance minima depends essentiallyon the effective capacitance C eff and wire circumference P .Equation (6) is only valid for V g values for which E F is lo-cated in the bulk gap. We estimate this V g -range from the gate-voltage dependent resistance curves plotted in Fig. 3. Thecorresponding region is highlighted by the dashed horizon-tal lines in Fig. 6(b) and matches the region of longest phase-coherence lengths l φ discussed above.The key quantity which is missing for comparing Eq. (6)with the experimental data of Fig. 6(b) is the effective capac-itance C eff , which is not directly accessible from experiment.Therefore, we resort to a numerical solution of the Poissonequation (described in Sec. IV) and the definition of C eff givenin Sec. V C. The anticipated result of such a calculation is C eff = 3 . · − F/m , carried out for the geometry anddielectric constants of the present wire. The curves obtainedfrom Eq. (6) both for spin helical and trivial surface statesare shown in Fig. 6(b). In the gap region the experimentaldata points are best described by the model invoking spin-helical surface states (solid lines); trivial surface states leadto a steeper slope (red dashed line) and fail to describe the ex- (a) j (j+1/2)(cid:215) G ( e / h ) V g (V) (b) j (j+1/2)(cid:215) k f2 = 4 n k f2 = 2 n k f2 = 4 n w/ P tra V g ( V ) N-N (subband index) E c E v Figure 6. (a) Conductance oscillations due to subband quantiza-tion. Conductance corrections ∆ G as a function of gate voltage forinteger (blue curve) and half-integer (orange curve) values of φ/φ ,obtained by averaging several line-cuts taken along constant φ fromthe data set of device w1 in Fig. 4(a). The curves exhibit a clearanti-correlated behavior. Each minimum-maximum pair (marked byvertical dashed lines) is assigned an index N that can be associatedwith a subband of the wire, see text. (b) Gate voltage position ofconductance maxima from (a) as a function of (subband) index N (with arbitrary offset N ). Within the bulk gap (marked by dashedhorizontal lines) the red curve segments denote expected behavior for V g ( N ) based on models for helical Dirac-type (solid line) and trivialspin-degenerate (dashed) surface states, see main text. Agreementof the conductance data with the former model implies the topolog-ical character of the surface states. For comparison, the blue lineshows corresponding results for a Dirac surface state model, basedon a trapezoidal cross section instead of a simple rectangular crosssection. The grey area around the data points is obtained repeatingthe analysis while introducing an artificial error of up to ± . in φ ,showing that the method is robust against errors in φ . periment. It would be interesting to apply this kind of analysisto wires with trivial surface states .The analysis described in this section was carried out forfive devices with different cross sections A . For each de-vice the nanowire circumference P , which defines the sub-band splitting ∆ k l = 2 π/P , was extracted from the SEMmicrographs. Alternatively, one can extract the circumferenceby fitting the data points, shown for samples w1 - w5 in theinset of Fig. 7, with Eq. (6). Using the corresponding effectivecapacitances C eff , the only remaining fit parameter is ∆ k l, fit . V g ( V ) N-N w1 w2 w3 w4 w5 k l , f i t ( / ¯ ) k l (1/¯) Figure 7. Analysis of subband spacing. The inset shows the de-pendencies V g ( N − N ) of samples w1-w5. The black dashed linesare best fits (within the bulk band gap) based on Eq. (6) with ∆ k l as the only fit parameter. Main panel: Best fitting values ∆ k l, fit ofsamples w1-w5 are plotted versus ∆ k l = 2 π/P , determined fromthe nanowires’ circumference P taken from SEM micrographs. Thedashed straight line marks the condition ∆ k l = ∆ k l, fit . Deviceswhose data is presented in this article are marked by squares. The best fits in the relevant V g -ranges for all five investigatedsamples are shown in the inset of Fig. 7. In the main panel ofFig. 7 the extracted values of ∆ k l, fit are plotted versus ∆ k l .The plot shows that ∆ k l ≈ ∆ k l, fit within experimental accu-racy, thus confirming the suitability of the analysis.In the following we present an in-depth study of thenanowire electrostatics, band structure and magnetoconduc-tance, that provides the theoretical basis for the analysis ofFig. 6 indicating the existence of helical surfaces states. IV. ELECTROSTATICS OF WIRE GEOMETRY
Since the capacitance used in the preceding section is notexperimentally accessible we need to resort to numerical anal-ysis. Furthermore, in the experimental setup, the Fermi en-ergy is locally tuned by a gate electrode that covers roughlythe upper part of the nanowire [see Fig. 1(c)] and induces anon-uniform surface electron density n ( z, s ) , where z is thelongitudinal ( < z < L ) and s the circumferential coordi-nate ( < s < P = 2 w + 2 h ). To account for the inhomoge-neous charge density in our subsequent transport simulationswhile keeping the model as simple as possible, we assumethat n ( z, s ) = n ( s ) is constant along the z -direction. We im-plemented the finite-element based partial-differential equa-tion (PDE) solver FE NI CS combined with the mesh gen-erator GMSH to obtain the gate capacitance, considering a2D electrostatic model for the geometry sketched in Fig. 1(c).Furthermore, the HgTe wire along with the Au top gate areboth assumed perfectly metallic with vanishing electric fieldin the interior, which implies Dirichlet-type boundary condi-tions on the corresponding surfaces, i.e. the electric poten-tial u ( x, y ) = V g on the boundary of the Au top gate and u ( x, y ) = 0 on the HgTe nanowire.The PDE solver numerically yields solutions of the Laplaceequation ∇ u ( x, y ) = 0 for the heterostructure, an examplewith V g = 1V is given in Fig. 8(a). The induced surfacecharge density n ( s ) is then given by the gradient of the electricpotential at the nanowire surface according to n ( s ) = ( (cid:15) r (cid:15) /e )( ∇ u ) · ˆ e n , (7)where ˆ e n is the unit vector defining the surface normal, (cid:15) r isthe dielectric constant of the insulating layer surrounding thenanowire ( (cid:15) r = 3 . for SiO and (cid:15) r = 13 . for HgCdTe) and (cid:15) is the vacuum permittivity. The gate capacitance per unitcharge, C ( s ) /e , defined as the surface electron density n ( s ) per gate voltage V g , is then exported from the PDE solver. Thesurface charge density at arbitrary gate voltages is obtained byexploiting the linearity n ( s ) = [ C ( s ) /e ] V g without the needfor repeating the electrostatic simulation.Figure 8(b) shows the capacitance C ( s ) resulting from thesimulation. The large spikes in the capacitance stem from thesharp edges of the HgTe nanowire. Electrostatic simulationsfor a trapezoidal wire cross-section with smoother gate pro-files (not shown) show that our simple model overestimatesthe effective capacitance by < . Assuming metallic surfacestates and ignoring thus the quantum capacitance also leads tocapacitance values which are larger. The finite length ofour wires which is ignored in our 2D electrostatic model un-derestimates the capacitance by the fringe fields at the ends.By analogy with finite length cylindrical capacitors the errorfor our wires is estimated to be ≤ . As the errors dueto the idealizations within our model tend to compensate eachother, we expect the calculated values to be quite accurate. V. BANDSTRUCTURE OF GATED NANOWIREA. Dirac surface Hamiltonian
Assuming that the bulk is insulating we model the HgTenanowire by means of the Dirac surface Hamiltonian H = v F [ p z σ z + ( p s + eA s ) σ y ] , (8)where A s is the vector potential which creates the longitu-dinal magnetic field. The numerical results presented in thefollowing were obtained using the python software packageKwant . Since Kwant utilizes tight-binding models, Eq. (8)needs to be discretized leading to Fermion doubling . Inorder to circumvent this, we add a small term quadratic in themomentum to the Hamiltonian which removes the artificialvalleys from the considered energy range (for a recent discus-sion see ).The magnetic flux and the curvature-induced Berry phaseare implemented via the boundary condition Ψ( z, s + P ) = Ψ( z, s )e i π ( φ/φ +1 / . (9) A BD C x [nm] y [ n m ] a)b) A B C D A c) s [nm]
Figure 8. Electrostatics of gated nanowires. (a) Electric poten-tial u ( x, y ) across the heterostructure of device w2. (b) Capacitanceprofile C ( s ) along the circumference of nanowire w2 obtained viaEq. (7) from the electric potential depicted in (a). (c) Capacitanceprofile for the simplified model (see text) for parameters P = 460 nm and C mean = 5 × − F m − . We further account for the effect of the top gate by adding theonsite energy (see corresponding Eq. (4) with g s = 1 ) E gate ( s ) = − (cid:126) v F (cid:113) πV g C ( s ) /e , (10)which induces the correct charge density n ( s ) along thenanowire circumference. B. Gate effect: Simplified capacitance model
In order to illustrate the effect of the gate-induced poten-tial on the bandstructure of a nanowire, we start with a sim- x x a)c) b) k z [nm -1 ] k z [nm -1 ] k z [nm -1 ] Figure 9. Gate-dependent splitting of Dirac-type nanowire bandstructure for simplified step capacitance model for top and bottom surface[see Fig. 8(c)]. (a) Sketch of the bandstructure for zero (left) and finite (right) gate voltage. Subband bottoms at k z = 0 are indicated by bluerectangular boxes. (b) Probability distributions of two representative states along wire circumference with capacitances C top and C bot at topand bottom surface. The state in the upper (lower) panel corresponds to an energy marked by a blue (yellow) cross in panel (c). (c) Calculatedbandstructures for three magnetic fluxes φ/φ = 0 , / , / for V g = 1 V and v F = 5 × m / s . In the left panel the dashed blackhorizontal line marks the Fermi level and the dashed yellow lines positions of shifted Dirac points.. In the middle panel areas correspondingto flux-sensitive (insensitive) energy levels are marked by dark (light) green color. ple step-shaped capacitance [see Fig. 8(c)] before examiningthe more realistic case shown in Fig. 8(b). In this simplifiedmodel, C ( s ) is determined by two capacitance values, onefor the top surface C top and one for the bottom surface C bot ,neglecting separate profiles at the narrow side surfaces. Wechoose C top /C bot = 5 for didactic purposes; in the experi-ments the ratio is ≈ .In the following, we first use a sketch of the resulting band-structure in Fig. 9(a) to explain the mechanisms that lead tothe corresponding numerical results shown in Fig. 9(c). For V g = 0 the bandstructure is given by a simple 1D Dirac cone[left panel in Fig. 9(a)] with quantized subbands owing to thefinite circumference. The positions of the subband minimaat k z = 0 are marked by blue rectangular boxes. The fluxthrough the nanowire is chosen to be . φ implying a stateat zero energy (marked by a red box). For V g > the Diraccone splits, so that the distances between the common Fermilevel E F and the two Dirac points are E top = (cid:126) v F (cid:112) πn top and E bot = (cid:126) v F √ πn bot [right panel in Fig. 9(a)]. Thedifference arises since the top surface is filled faster thanthe bottom one. The Fermi energy is given by the average E F = ( E top + E bot ) / (cid:126) v F (cid:112) πV g /e (cid:0)(cid:112) C top + √ C bot (cid:1) .Interestingly, the splitting of the Dirac cone does not in-fluence the k z = 0 subband spacing, which is perfectly pre-served. Furthermore, for k z (cid:54) = 0 the states can be divided intotwo groups. States with energies in the dark green regions in Fig. 9 extend over the entire wire circumference and conse-quently are flux-sensitive, whereas states with energies in thelight green regions are localized on the upper or lower surfaceand hence are not susceptible to flux changes. Representativeexamples of both kinds of states are shown in Fig. 9(b). Thereason for this peculiar behavior is Klein tunneling . Modesat k z = 0 perpendicularly hit the potential step associatedwith the capacitance profile of Fig. 8(c) and are thus unaf-fected by its presence – Klein tunneling is perfect. On theother hand, for k z (cid:54) = 0 the light green regions host states fromone Dirac cone only, the dark green ones from both. Since k z is conserved during tunneling, it is only in the dark greenregions that electrons can Klein-tunnel from one cone to theother, yielding hybridized extended states.The different flux sensitivity of the two classes of statesis also reflected in the numerical bandstructures shown inFig. 9(c) for three different fluxes: While the hyperbolic en-ergy levels in the regimes corresponding to the light greenareas are identical (on scales resolved in the figures) for allthree fluxes, the energy levels belonging to the dark green ar-eas obviously change with varying flux. We note in passingthat trivial surface states which might form at the etched sidesurfaces do not contribute to the oscillations and thus to ouranalysis as they are localized at the sides. C. Gate effect: realistic capacitance model
We now use these insights to examine the results obtainedwith the more realistic capacitance profile shown in Fig. 8(b).We present results for the nanowire geometry of sample w2;the other geometries do not qualitatively change the results.The corresponding bandstructures are presented in Fig. 10 forthree different magnetic fluxes and V g = 0 . V. The inter-pretation of the bandstructure is not as simple as in the stepcapacitance model, as there are now four nanowire surfacesinvolved, each with a more complicated capacitance profile.Furthermore, the difference between the “potential bottoms”on each surface is smaller. However, the main features dis-cussed in Sec. V B are still present, since they are actuallyindependent of the capacitance profile shape.Notably, Fig. 10 shows that the k z = 0 subband spacingis perfectly preserved, even though the transversal wave vec-tor k l ( s ) = (cid:112) πn ( s ) is now s -dependent. Indeed, we cangeneralize the discussion of the simplified step capacitancemodel: For k z = 0 the motion is purely angular, and thusKlein tunneling is perfect irrespective of the complexities ofthe potential profile. Hence, the electron wave function expe-riences an average of the gate potential, whose correspondingwave vector average fulfils (cid:104) k l ( s ) (cid:105) ≡ P P (cid:90) d s k l ( s ) = 2 πP ( l + 0 . − φ/φ ) . (11)This leads to a gate-independent subband spacing ∆ k l =2 π/P that can be also derived by solving the Dirac equationfor k z = 0 . Moreover, in the experimentally relevant parame-ter range, the Fermi level stays always above both Dirac pointsfor V g > , and the subband minima at E F are always at k z = 0 .For a general capacitance profile C ( s ) the Fermi energy isgiven by the average E F = (cid:126) v F (cid:113) πV g /e (cid:68)(cid:112) C ( s ) (cid:69) = (cid:126) v F (cid:113) πV g C eff /e, (12)where the effective value C eff ≡ (cid:68)(cid:112) C ( s ) (cid:69) enters instead ofthe mean C mean ≡ (cid:104) C ( s ) (cid:105) with (cid:104) . . . (cid:105) denoting the mean valuealong the circumference. The value for C eff calculated fromthe capacitance profile of w1 was used in the analysis of theexperimental data in Fig. 6. Note that the difference between C eff and C mean is given by the variance of the capacitanceprofile Var( (cid:112) C ( s )) = C mean − C eff . For the step capacitancemodel with C top = 5 C bot this difference is clearly visible[ C eff /C mean ≈ . , see Fig. 8(c)] but in the experiments itis typically quite small (e.g. for w2 C eff /C mean ≈ . ).As will become evident in the following, the perfect sub-band quantization and the fact that the subband minima arelocated at k z = 0 are crucial ingredients to probe the signa-ture of the Dirac surface states in transport. VI. CONDUCTANCE SIMULATIONS
The conductance simulations presented in the followingwere carried out by using an extended version of the tight-binding model introduced in Sec. V A. The extended systemincludes coupling of the nanowire to leads with the same ge-ometry as the nanowire but with negative onsite energy (i.e.highly-doped leads) to account for the wide leads in the ex-periment. Residual disorder in the wires used in experiment ismodelled by adding a random disorder potential V ( r ) definedthrough the correlator (cid:104) V ( r ) V ( r (cid:48) ) (cid:105) = K ( (cid:126) v F ) πξ e −| r − r (cid:48) | / ξ . (13)Here, ξ is the correlation length and K determines the disor-der strength .Figure 11 shows the conductance G ( V g ) for φ = 0 and φ = 0 . φ . Both curves exhibit distinct oscillations on topof an increasing conductance background (similar to corre-sponding calculations in Ref. ). The anti-correlated behaviorof the two oscillatory curves is due to the flux-sensitivity of thestates near k z = 0 . The oscillatory behavior can be explainedby the bandstructure: Whenever the Fermi energy approachesthe bottom of one of the disorder-broadened subbands, thehigh density of states (associated with a van-Hove singular-ity) causes enhanced scattering and thus leads to a reductionof the conductance. By further increasing the gate voltage,the additional conductance channel fully opens and the Fermienergy leaves the vicinity of the van-Hove singularity, both ef-fects leading to an increasing conductance. Thus, the conduc-tance oscillation, related to successive subband opening, is afingerprint of the bandstructure at k z = 0 , as also observed inthe transport measurements. Due to the undisturbed subbandquantization at k z = 0 , the distance between two conductanceminima (one oscillation period) corresponds to a change of ∆ k l = 2 π/P in the Fermi wave vector, as in the case with-out any inhomogeneity caused by the top gate. However, theinhomogeneity enters via C eff , which determines the Fermiwave vector k F = (cid:112) πV g C eff /e [see Eq. (12)]. This justifies a posteriori the validity of Eq. (6) and the procedure used inSec. III B to analyze the experimental data if the surface statesare Dirac-like.Trivial surface states, however, are expected to have aquadratic dispersion, i.e. Klein tunneling is absent. Thus, thequestion arises whether the ∆ G ( V g ) -oscillations would stillallow for the quantitative analysis performed in Sec. III B. Oursimulations (not shown) for a quadratic dispersion reveal thatfor realistic gate-induced potentials the subband minima arestill at k z = 0 but now states below E F can be confined andthus become flux-insensitive. Moreover, the subband quanti-zation ∆ k l = 2 π/P is no longer preserved. However, for E F in our experimental range the k z = 0 modes, which lead to theconductance oscillations, have an angular motion energy thatis much larger than the gate potential. These states are thusflux-sensitive and the subband quantization is only mildly af-fected (the degeneracy with respect to angular momentum islifted). The lifting of the degeneracy is small compared to0 k z [nm -1 ] k z [nm -1 ] k z [nm -1 ] Figure 10. Calculated bandstructures of device w2 based on the realistic capacitance model [see Fig. 8(b)] for V g = 0 . and three fluxes φ/φ = 0 , / , / (from left to right). The spacing between subband minima at k z = 0 is constant for given flux. extrema Figure 11. Calculated disorder-averaged conductance as a func-tion of gate voltage (starting from the Dirac point) for zero flux (bluecurve) and half a flux quantum (red curve). Calculations are per-formed for the nanowire geometry and using C eff of sample w2. Theaverage is taken over disorder configurations, based on the im-purity model (Eq. (13)) with disorder strength K = 0 . and corre-lation length ξ = P/ . Minima-maxima pairs in the conductanceare marked with gray vertical lines. The inset shows the gate po-sition of the minima-maxima pairs as a function of subband index N , following the same evaluation as performed in Fig. 5(b) for theexperimental data. the subband spacing and cannot be observed within the ex-perimental precision. For trivial surface states, we thereforeindeed expect to measure a conductance with a similar shapeas for Dirac states but with minima-maxima pairs followingthe spin-degenerate version ( g s = 2 ) of Eq. (6).The inset of Fig. 11 shows a similar evaluation of theminima-maxima pairs as was done for the experimental datain Fig. 5(b). The green parabolic curve describing spin-helicalDirac states according to Eq. (6) (with g s = 1 ) matches per-fectly with the results (black bullets) from the analysis ofthe gate-dependent conductance extrema, whereas the purplecurve which holds for spin-degenerate trivial surface states( g s = 2 ) is way off. Since we used Eq. (10) (with g s = 1 )to simulate the gate effect, the agreement was essentially ex-pected. However, this analysis shows that if the surface states in 3D HgTe nanowires are Dirac-like, one should be able toobtain this signature of spin non-degenerate states by conduc-tance measurements, despite the complicated potential profileinduced by the top gate, as long as C eff is known. VII. CONCLUSIONS AND OUTLOOK
We fabricated nanowires based on strained HgTe and inves-tigated in detail their peculiar transport properties in a joint ex-perimental and theoretical effort. With regard to topologicalinsulator properties, HgTe-based systems represent an inter-esting alternative to Bi-based systems, as surface states in theformer appear to be well decoupled from bulk states, and addi-tionally feature high surface mobilities . The nanowires werebuilt out of strained bulk systems in a well controlled way.In particular, we demonstrated that in these mesoscale con-ductors the topological properties of the corresponding bulksystems prevail, and appear in combination with quantum co-herent effects: The observed h/e -periodic Aharonov-Bohm-type conductance modulations due to a coaxial flux clearlyindicate that transport along the wires is indeed both surface-mediated and quasi-ballistic, and additionally phase-coherentat micron scales. At low temperatures we moreover foundthat the extracted phase-coherence lengths are increased up to5 µ m upon tuning the wire Fermi energy into the bulk bandgap, were topological surface transport is singled out.Besides the Aharonov-Bohm oscillations, we observed andexamined in detail further distinct conductance oscillationsappearing as a function of a gate voltage. We showed thatthe spacing of the observed regular gate-dependent oscil-lations reveals the topological nature of the surface states:The gate dependence is only compatible with a model as-suming non-degenerate (Dirac-type helical) surface states,and rules out usual spin-degenerate states. This identifica-tion required on the theory side a quantitative electrostaticcalculation of the gate-induced inhomogenous charge car-rier density and associated capacitance of the whole (wireplus gate) system. The latter entered the evaluation of thetransport data, and was furthermore integrated into numeri-cal tight-binding magneto-transport calculations to show thatthe gate-dependent conductance oscillations obtained indeed1agree with experiment. This theoretical analysis also shedlight on the role of Klein tunneling for the flux-sensitive sur-face states extending around the entire wire circumference,which govern the wire bandstructure.Our finding that spin non-degenerate states exist on the wiresurface suggests in particular that for φ/φ = 0 . , wheretime reversal symmetry is restored, the total number of left- orright-moving states is always odd at arbitrary Fermi level, im-plying a topologically protected perfectly transmitted mode.In the presence of an s -wave superconductor, which opens agap in the nanowire bandstructure via the proximity effect,Majorana fermions are expected to appear at the endings of the3DTI wire, if the latter hosts such an odd number of states at E F . This opens up the interesting possibility of switchingfrom a topologically trivial (even number of states at E F ) toa non-trivial situation (odd number of states at E F ) by addinghalf a flux quantum through the wire’s cross section. This no-tably led to a recent proposal to use proximitized TI nanowiresas building blocks for (coupled) topological Majorana qubitsand networks of those . While such proposals usually as-sume a uniform carrier density around the wire circumfer-ence, in experiments such density is expected to be stronglyinhomogeneous, especially if the Fermi level is tuned via a top/back gate voltage. This raises the question of how sub-stantially the topological behavior would be affected by stronginhomogeneities. Here we demonstrated through the quanti-tative analysis of our data that the essential characteristics oftopological transport are indeed preserved under realistic ex-perimental conditions.To conclude, our work puts forward HgTe-based topolog-ical insulator nanowires as a promising and realistic plat-form for exploring a wealth of phenomena based on spin-momentum locked quantum transport and topological super-conductivity. ACKNOWLEDGMENTS
This work was supported by
Deutsche Forschungsgemein-schaft (within Priority Program SPP 1666 ”Topological Insu-lators” ) and the
ENB Doktorandenkolleg ”Topological Insu-lators” . Support by RF President Grant No. MK-3603.2017.2and RFBR Grant No. 17-42-543336 is also acknowledged.We thank J. Bardarson, J. Dufouleur, S. Essert and E. Xypakisfor useful conversations. N. Z. Hasan and C. K. Kane, “
Colloquium : Topological insula-tors,” Rev. Mod. Phys. , 3045 (2010). J. H. Bardarson and J. E. Moore, “Quantum interference andAharonov-Bohm oscillations in topological insulators,” Rep.Prog. Phys. , 056501 (2013). V. E. Sacksteder IV, K. B. Arnardottir, S. Kettemann, andI. A. Shelykh, “Topological effects on the magnetoconductivityin topological insulators,” Phys. Rev. B , 253148 (2014). H. Peng, K. Lai, D. Kong, S. Meister, Y. Chen, X.-L. Qi, S.-C.Zhang, Z.-X. Shen, and Y. Cui, “Aharonov-Bohm interference intopological insulator nanoribbons,” Nat. Mater. , 225 (2010). F. Xiu, L. He, Y. Wang, L. Cheng, L.-T. Chang, M Lang,G. Huang, X. Kou, Y. Zhou, X. Jiang, Z. Chen, J. Zou, A. Shai-los, and K. L. Wang, “Manipulating surface states in topologicalinsulator nanoribbons,” Nat Nano , 216 (2011). P. M. Ostrovsky, I. V. Gornyi, and A. D. Mirlin, “Interaction-Induced Criticality in Z Topological Insulators,” Phys. Rev. Lett. , 036803 (2010). J. H. Bardarson, P. W. Brouwer, and J. E. Moore,“Aharonov-Bohm Oscillations in Disordered Topological Insula-tor Nanowires,” Phys. Rev. Lett. , 156803 (2010). Y. Zhang and A. Vishwanath, “Anomalous Aharonov-Bohm Con-ductance Oscillations from Topological Insulator Surface States,”Phys. Rev. Lett. , 206601 (2010). G. Rosenberg, H.-M. Guo, and M. Franz, “Wormhole effect in astrong topological insulator,” Phys. Rev. B , 041104 (2010). T. Ando and H. Suzuura, “Presence of Perfectly ConductingChannel in Metallic Carbon Nanotubes,” Jour. Phys. Soc. Japan , 2753 (2002). M. Tian, W. Ning, Z. Qu, H. Du, J. Wang, and Y. Zhang, “Dualevidence of surface Dirac states in thin cylindrical topological in-sulator Bi Te nanowires,” Scientific reports , 1212 (2013). J. Dufouleur, L. Veyrat, A. Teichgräber, S. Neuhaus, C. Nowka,S. Hampel, J. Cayssol, J. Schumann, B. Eichler, O. G. Schmidt, B. Büchner, and R. Giraud, “Quasiballistic Transport of DiracFermions in a Bi Se Nanowire,” Phys. Rev. Lett. , 186806(2013). M. Safdar, Q. Wang, M. Mirza, Z. Wang, K. Xu, and J. He, “Topo-logical Surface Transport Properties of Single-Crystalline SnTeNanowire,” Nano Letters , 5344 (2013). S. S. Hong, Y. Zhang, J. J. Cha, X.-L. Qi, and Y. Cui, “One-Dimensional Helical Transport in Topological Insulator NanowireInterferometers,” Nano Letters , 2815 (2014). S. Cho, B. Dellabetta, R. Zhong, J. Schneeloch, T. Liu, G. Gu,M. J. Gilbert, and N. Mason, “Aharonov-Bohm oscillations in aquasi-ballistic three-dimensional topological insulator nanowire,”Nat Commun , 7634 (2015). L. A. Jauregui, M. T. Pettes, L. P. Rokhinson, L. Shi, and Y. P.Chen, “Magnetic field-induced helical mode and topological tran-sitions in a topological insulator nanoribbon,” Nature Nanotech-nology , 345– (2016). J. Kim, A. Hwang, S.-H. Lee, S.-H. Jhi, S. Lee, Y. C. Park, S. Kim,H.-S. Kim, Y.-J. Doh, J. Kim, and B. Kim, “Quantum ElectronicTransport of Topological Surface States in β -Ag Se Nanowire,”ACS Nano , 3936 (2016). C. Brüne, C. X. Liu, E. G. Novik, E. M. Hankiewicz, H. Buh-mann, Y. L. Chen, X. L. Qi, Z. X. Shen, S. C. Zhang, and L. W.Molenkamp, “Quantum Hall Effect from the Topological Sur-face States of Strained Bulk HgTe,” Phys. Rev. Lett. , 126803(2011). O. Crauste, Y. Ohtsubo, P. Ballet, P. A. L. Delplace, D. Carpen-tier, C. Bouvier, T. Meunier, A. Taleb-Ibrahimi, and L. Lévy,“Topological surface states of strained Mercury-Telluride probedby ARPES,” arXiv preprint arXiv:1307.2008 (2013). S.-C. Wu, B. Yan, and C. Felser, “Ab initio study of topologicalsurface states of strained HgTe,” EPL (Europhysics Letters) ,57006 (2014). P. G. Silvestrov, P. W. Brouwer, and E. G. Mishchenko, “Spin and charge structure of the surface states in topological insulators,”Phys. Rev. B , 075302 (2012). K.-M. Dantscher, D. A. Kozlov, P. Olbrich, C. Zoth, P. Fal-termeier, M. Lindner, G. V. Budkin, S. A. Tarasenko, V. V.Bel’kov, Z. D. Kvon, N. N. Mikhailov, S. A. Dvoretsky, D. Weiss,B. Jenichen, and S. D. Ganichev, “Cyclotron-resonance-assistedphotocurrents in surface states of a three-dimensional topologicalinsulator based on a strained high-mobility HgTe film,” Phys. Rev.B , 165314 (2015). D. A. Kozlov, Z. D. Kvon, E. B. Olshanetsky, N. N. Mikhailov,S. A. Dvoretsky, and D. Weiss, “Transport Properties of a 3DTopological Insulator based on a Strained High-Mobility HgTeFilm,” Phys. Rev. Lett. , 196801 (2014). A. G. Aronov and Yu. V. Sharvin, “Magnetic flux effects in disor-dered conductors,” Rev. Mod. Phys. , 755 (1987). S. Washburn and R. A. Webb, “Aharonov-Bohm effect in normalmetal quantum coherence and transport,” Advances in Physics ,375 (1986). D. A. Kozlov, D. Bauer, J. Ziegler, R. Fischer, M. L. Savchenko,Z. D. Kvon, N. N. Mikhailov, S. A. Dvoretsky, and D. Weiss,“Probing Quantum Capacitance in a 3D Topological Insulator,”Phys. Rev. Lett. , 166802 (2016). A. Savitzky and M. J. E. Golay, “Smoothing and Differentiation ofData by Simplified Least Squares Procedures,” Analytical Chem-istry , 1627 (1964). Ö. Gül, N. Demarina, C. Blömers, T. Rieger, H. Lüth, M. I. Lepsa,D. Grützmacher, and Th. Schäpers, “Flux periodic magnetocon-ductance oscillations in GaAs/InAs core/shell nanowires,” Physi-cal Review B (2014), 10.1103/physrevb.89.045417. A. Logg, K.-A. Mardal, G. N. Wells, et al. , Automated Solutionof Differential Equations by the Finite Element Method (Springer,2012). C. Geuzaine and J.-F. Remacle, “Gmsh: a three-dimensional fi-nite element mesh generator with built-in pre- and post-processingfacilities,” International Journal for Numerical Methods in Engi-neering (2009). M.-H. Liu, “Theory of carrier density in multigated dopedgraphene sheets with quantum correction,” Phys. Rev. B ,125427 (2013). B.N. Das and S.B. Chakrabarty, “Capacitance and charge distri-bution of two cylindrical conductors of finite length,” IEE Pro-ceedings - Science, Measurement and Technology , 280–286(1997). C. W. Groth, M. Wimmer, A. R. Akhmerov, and X. Waintal,“Kwant: A software package for quantum transport,” New Journalof Physics , 1 (2014). L. Susskind, “Lattice fermions,” Physical Review D , 3031(1977). R. Stacey, “Eliminating lattice fermion doubling,” Physical Re-view D , 468 (1982). K. M. M. Habib, R. N. Sajjad, and A. W. Ghosh, “Modified DiracHamiltonian for efficient quantum mechanical simulations of mi-cron sized devices,” Applied Physics Letters , 113105 (2016). P. E. Allain and J. N. Fuchs, “Klein tunneling in graphene: opticswith massless electrons,” The European Physical Journal B ,301 (2011). For a detailed discussion of scattering in quasi-ballistic Diracnanowires, see . A. Cook and M. Franz, “Majorana fermions in a topological-insulator nanowire proximity-coupled to an s -wave superconduc-tor,” Phys. Rev. B , 201105 (2011). F. de Juan, R. Ilan, and J. H. Bardarson, “Robust Transport Sig-natures of Topological Superconductivity in Topological InsulatorNanowires,” Phys. Rev. Lett. , 107003 (2014). J. Manousakis, A. Altland, D. Bagrets, R. Egger, and Y. Ando,“Majorana qubits in a topological insulator nanoribbon architec-ture,” Phys. Rev. B , 165424 (2017).42