Arxiv:1306.2038V2 [Cond-Mat.Mtrl-Sci] 28 Aug 2013 Em Ee Ssﬃinl Ls Otetpo H Gap

Home , Bismuth selenide

arXiv:1306.2038v2 [cond-mat.mtrl-sci] 28 Aug 2013 em ee ssffiinl ls otetpo h gap. the of conduction bulk top by the produced the be to also when can close re- state it Alternatively, sufficiently a surface The is as the point. level TI from Dirac Fermi a arise state in may surface MR field its negative zero at the gap-opening near of predict sult MR theories negative recent or various WL materials, these describe transport systems. of these understanding in comprehensive oscillation more quantum a wards separate of accurately fields discovery effects to higher the at ability effects the has assignment, (WL) mobility) localization simple low weak Dirac their this of state challenged because surface (presumably bands the wide on from a cone. results locking that in (WAL) and spin-momentum antilocalization weak fields thicknesses. of magnetic film terms low of magnetore- at range positive ungated show (MR) films on sistance TI measurements uncompensated and transport and point QLs temperature band. Dirac conduction three the low samples bulk above uncompensated with the level made, Fermi intersecting As a (QLs), have cell. typically layers unit a quintuple forming atomic of magnetic image an to induce close to charge in- electric predicted full monopole. an is a and surface surface, opened, the the is on applied gap is sulating layer magnetic thin iesoa I(DT)hsbe bismuth-selenide been has (3D-TI) (Bi TI dimensional ftm eeslsmer,tesraesae antbe impurities. cannot states non-magnetic surface by Because the back-scattered symmetry, impurities. reversal non-magnetic time cannot any of that by states destroyed surface gapless be contains and bulk, the mrigWa oaiainEet nTplgclInsulato Topological on Effects Localization Weak Emerging iI ag( Yang I. Qi yaotn oooial o-rva aitnasto Hamiltonians non-trivial topologically adopting By rbbytems xesvl tde three- studied extensively most the Probably oooia nuao T)hsafl nrygpin gap energy full a has (TI) insulator topological A 2 Se 10 15 3 i Liu, Min , 13 ), nhg-ult lsmypoieafis tpto- step first a provide may films high-quality in 5 4 – 3 eateto hmclEgneigadMtrasScience, Materials and Engineering Chemical of Department hl h nblt oacutfrtebulk the for account to inability the While 7 , 4 nsa eaiemgeoeitnewsosre ertez the near observed ( was magnetoresistance negative Unusual infrftr elzto ftehl-nee unie a quantized half-integer the insulators. of topological realization future for tion rxmt ffc ewe oooia nuao n ninsu an and insulator topological a between effect proximity eoee above recovered than 杨 xiiigcytlsrcueta consists that structure crystal exhibiting T C hnfim ftplgclisltrBi insulator topological of films Thin 弃 ,2 1, ,rsmln h eklclzto ffc;weesteusua the whereas effect; localization weak the resembling ), t ), 2 6 eal aoaoyfrAvne aeil,Safr Unive Stanford Materials, Advanced for Laboratory Geballe ∼ lxne Palevski, Alexander ,2, 1, tnodNncaatrzto aoaoy tnodUniv Stanford Laboratory, Nanocharacterization Stanford 5 n,we t o n otmsrae r ope.Teeresu These coupled. are surfaces bottom and top its when 4nm, colo hsc n srnm,TlAi nvriy 69978 University, Aviv Tel Astronomy, and Physics of School ∗ 3 eateto ple hsc,Safr nvriy Stanf University, Stanford Physics, Applied of Department ea Dolev, Merav 10 1 eateto hsc,Safr nvriy tnod A9 CA Stanford, University, Stanford Physics, of Department T – C 12 uhngtv antrssac a nyosre o Bi for observed only was magnetoresistance negative Such . hswsepandin explained was This ,3 2, 5 8 , n .Marshall, F. Ann 9 iZhang, Li nparticular, In (Bi 1 , 2 12 2 Dtd uy1,2018) 19, July (Dated: hna When 2 Se Se , 14 3 ,3 2, 3 and eedpstddrcl nisltn ermgei EuS. ferromagnetic insulating on directly deposited were ES Interface -EuS) 16 ifn Zhao, Jinfeng 6 uhs .Risbud, H. Subhash bqiosyi iia ls below films, similar in ubiquitously ermge I,ES.WieaoeteCretempera- Curie the above ( ture While EuS). (IF, ferromagnet in(fig. tion et(AE n te plctos esuidtein- (Bi the TI studied film thin we to- a applications, step between terface other initial and an ef- (QAHE) Hall as fect anomalous and quantized half-integer materials, realizing TI wards of state surface eotdi ae fgtdutati Bi ultra-thin Bi gated doped magnetically of cases in reported ob 15 be to QAHE. a with studied interferometer. Sagnac (SQUID) a were and device magnetometer films interference the quantum superconducting of properties Magnetic conduction surface the when suppressed. observed sufficiently be is only can and eiain eeosre ohv napeibeper- appreciable an have (fig. component to pendicular observed were netizations eie o 2K for verified 0 with resolution, interferometer tial scanning-Sagnac a ing hto lsgono o-ptxa usrtssc as such substrates to non-epitaxial compares on that grown Al quality films with of layers that top as grown then 0n eegono Al on well- 20– grown the thicknesses were of with method. 200nm films EuS (PLD) thin ferromagnet quality insulating deposition high studied layer, laser bottom pulsed the For a using grown IF. the and TI proximity the a between indicating clearly effect field, zero the near negative h lswscnre ihXrydffato (XRD). diffraction X-ray with confirmed better was films indicated the (AFM) 2 microscopy pow- force than EuS Atomic purity high of der. (SPS) sintering plasma spark by ofrhreuiaeteuiuns ftasoti the in transport of uniqueness the elucidate further To l iae B)smlspeetdi hswr were work this in presented samples (BL) bilayer All 2 O oaosHl ffc nthree-dimensional in effect Hall nomalous 3 T hl oedtiso rpriso PLD-grown of properties on details more While . C ufc mohes n h rsalnt of crystallinity the and smoothness, surface A ˚ nvriyo aiona ai,C 95616 CA Davis, California, of University 4 aigfroant aigtefounda- the laying ferromagnet, lating T fteESw n oiieM hti observed is that MR positive find we EuS the of ) . r edblwteCretemperature Curie the below field ero K oa ereeto h lsa 0K us- K, 10 at films the of effect Kerr Local 7K. C lxne .Fried, D. Alexander 1e setmtdfo ui-es t(fig. fit Curie-Weiss a from estimated is .Eeti nuain( insulation Electric ). riy tnod A94305 CA Stanford, ersity, oiiemgeoeitnewas magnetoresistance positive l st,Safr,C 94305 CA Stanford, rsity, 22 < T < , 23 r,C 94305 CA ord, 4 niaehgl nfr magnetiza- uniform highly indicate t rvd vdnefra for evidence provide lts e vv Israel Aviv, Tel –nuaigFerromagnet r–Insulating n hrnKapitulnik Aharon and 2 Se 4305 0K Bi 300K. 2 17 3 2 O Se , films. 1a 18 3 3 ,2 1, srqie oraiethe realize to required as ) 00)uigtresmade targets using (0001) aesthinner layers uhapeoeo was phenomenon a Such 20 2 lzbt Schemm, Elizabeth Se 2 T Se 3 R C n ninsulating an and ) 3 h Rbecomes MR the 2 Te hnfim were films thin > 22 3 0Ω was 20MΩ) films ,2 3 2, 1, hi mag- Their . 9 µ spa- m 19 and ,2 1, 1b 21 ) 2

2 3 (a) (b) ples with diﬀerent thicknesses for the Bi2Se3 layer were 1 fabricated and transport properties of two representa- 2 TC = 15.7K tive thicknesses are shown in ﬁg. 2. As ubiquitously 0 (arb. unit) (arb. unit) z z 1 M M -1 H (T / µ ) H (T / µ ) H ^z 0 0 H = 0.02T/µ0 C = 0.023T/µ0 -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8 -2 0 0.15 0.3 -0.2 -0.1 0 0.1 0.2 5 10 15 20 25 (a) TI0: t ≈ 5QL (b) TI3: t ≈ 3QL H T 2K (T/µ0) (K) 2K 0.1 3K 0.2 (c) (d) 4K Bi Se EuS 5K 2 3 4K 6K Bi Se 2 3 Titanium 0.05 10K 7K 0.1 R(H) / R(0) 8K R(H) / R(0) ∆ 9K ∆ 20K 10K EuS 0 12K 0 0.2 14K 0.3 (c) BL0: t ≈ 5QL (d) BL3: t ≈ 3QL 16K z 18K y 1QL 0.15 2K 20K 0.2 x Al2O3 substrate 22K 4K 25K 0.1 30K 300 0 10K 0.1

10G R(H) / R(0) R(H) / R(0) (e) ∆ 0.05 ∆ 200 48G 40G 20K 0 30G 0

rad) 100

µ 20G ( -8 -6 -4 -2 0 2 4 6 8 -8 -6 -4 -2 0 2 4 6 8

¥ 4G H (T / µ ) 0 0 H (T / µ0) ¦ 10G Kerr

§ 20G

¤ 100 ¨ 30G FIG. 2. (Color online) Magnetoresistance (MR) and its tem-

200 48G perature dependence of PLD-grown Bi2Se3 thin ﬁlms (TI)

¢ ¡ 3 2 1 0 1 2 3 x / 45µm and Bi2Se3-EuS bilayer (BL) devices. (a)(b) The Bi2Se3-only samples have positive MR from WAL regardless of thick- FIG. 1. (Color online) (a) Magnetization of a EuS thin film nesses. (c)(d) The BL devices with TI-layer thicknesses t in perpendicular magnetic fields and (b) its dependence on & 4QL behave similarly to the TI-only films, whereas with t temperature. The black solid curve is a fitting to Curie-Weiss TI-layer thicknesses . 4QL a distinctive WL-like negative Law. (c) Schematic of a bilayer device: an EuS film was MR is observed at low-fields below the Curie temperature of EuS. The thickness limit coincides with occurrence of cou- deposited on an Al2O3(0001) substrate, on top of it 15nm Titanium contacts with gradual height on the edges and a pling between the top and the bottom surfaces of a TI thin film. Bi2Se3 layer were sequentially deposited. (d) TEM image of the cross-section of a bilayer device, showing the quintuple layers (QL) of Bi2Se3 and a smooth TI-IF interface. (e) Kerr seen in TI thin films, the PLD-grown Bi2Se3-only sam- Angles measured at 10K by a scanning Sagnac interferometer ples show positive MR at low fields regardless of thick- across the edge of the Bi2Se3 layer, where x< 0 region is bare nesses, which broadens monotonically with increasing EuS and x> 0 is Bi2Se3 on top of EuS. temperature (figs. 2a & 2b), consistent with weak antilocalization (WAL). Fittings to the standard Hikami- Larkin-Nagaoka (HLN) formula describing the WAL 21,24 13,25,26 Bi2Se3 and EuS films will be given elsewhere, we magnetoconductance yield dephasing lengths (lφ) note here that the important features necessary for a re- comparable to molecular-beam epitaxy (MBE)-grown liable study of the TI-IF proximity effect, which are an samples with the same thicknesses.11,14 For thicknesses of insulating ferromagnet with a perpendicular anisotropy the Bi2Se3 layer greater than ∼ 4QL, the TI-IF bilayers component, and a well defined interface to the TI mate- have similar low-field MR features to their TI-only coun- rial, are met in all our films. Specific for the transport terparts (fig. 2c). By contrast, for thicknesses t . 4QL, measurements, 15nm-thick Titanium was evaporated on the bilayers show distinctive negative low-field MR at low corners of the device to serve as Ohmic contacts (fig. 1c). temperatures (fig. 2d), resembling WL effects. Such neg- For shaping the layers, we used shadow masks to produce ative MR features are clearly distinguishable well below gradual height profiles and smooth overlapping near the the Curie temperature (TC = 15.7K) of the IF (figs. 3c & edges of Titanium contacts. The effectiveness was con- 3b), whereas positive MR from WAL appears at higher firmed with scanning electron microscopy (SEM). The fields. Below and close to TC (fig. 3b), the negative MR cross-section transmission electron micrographs (fig. 1d) can no longer be directly observed. However, its rem- indicate smooth interface and excellent layering of the nant contribution reverses the thermal broadening of the Bi2Se3. For comparison, Bi2Se3-only samples were made overall positive MR. This suggests that the WL effect in each of the PLD sessions in which the Bi2Se3 of BLs is reduced rapidly close to the ferromagnetic transition. were deposited, with Ohmic contacts between Bi2Se3 and Above TC (fig. 3a), the positive MR is eventually broad- bare Al2O3 (0001) substrates, i.e. identical configuration ened when increasing the temperature, similar to com- to fig. 1c except for the absence of an EuS layer. mon WAL features in TI-only thin films. The agreement To examine the relevance of surface conduction, sam- between TC and the temperatures at which WL becomes 3

µ µ H (T / 0) H (T / 0) -1.5 -1 -0.5 0 0.5 1 1.5 -1.5 -1 -0.5 0 0.5 1 1.5 thinner and therefore more resistive parts of the sample. Indeed, the sheet resistance of the BL samples are 1 1

-2 (a) (b) -2 one order of magnitude higher than that of the Bi2Se3- 0.8 0.8

/ 10 / 10 only samples with similar thicknesses (ﬁg. 4b). When 0.6 0.6 0.4 30K 16K 0.4 25K 14K 2 0.2 14 0 BL1 10 0.2 22K 12K 30° 20K 10K -2 12 0° (b) BL2 R(H) / R(0) 0 R(H) / R(0) 60° 45° ∆ 0 18K 9K ∆ 10 BL3 ° -1 ) / 10 16K 8K 70 2

-0.2 BL4 e 8 -1 0 1 1 / 8K 0.32K TI3 10 h

6 ° ( 4 (c) 7K (d) 0.4K 0.1 89 0.6K ❑ -2

6K 4 R 2 5K 0.8K R(H)/ R(0) / 10 1.1K 2 4K 0.05 ∆ 0 3K 0 (a) 10 0 2K 1.5K -2

R(H) / R(0) -8 -6 -4 -2 0 2 4 6 8 2 20 200

0 ∆ -2 H (T/µ0) R(H) / R(0) T (K)

∆ -4 2K -0.05 -1.5 -1 -0.5 0 0.5 1 1.5 -1 -0.5 0 0.5 1 FIG. 4. (Color online) (a) The WL and WAL features in BL4 µ µ H (T / 0) H (T / 0) broaden with increasing angles between the magnetic field and the normal of the sample plane, confirming their orbital FIG. 3. (Color online) Low-field MR of a TI-IF bilayer device origins. The insert is zoomed-in version of low-field features. (BL3, t ≈ 3QL) in magnetic fields perpendicular to the film: (b) Sheet resistances of the four TI-IF bilayer samples with (a) Above TC , a WAL-like positive MR sharpens with de- WL features (BL1–4) and a PLD-grown TI-only film (TI3) creasing temperature; (b) Just below TC , MR broadens with measured with the van der Pauw method. The non-vanishing decreasing temperature; (c) Well below TC , a WL-like nega- MR near parallel fields and the high resistances of the BL tive MR emerges near zero field. The emergence of WL below samples are consistent with the observed inhomogeneity in TC of the EuS indicates a clear TI-IF proximity effect. (d) MR sample thickness. of BL3 below T = 2K, with solid lines as illustrative guides. Resistance at T ≤ 0.8K was measured with two-terminal con- lowering the temperature, the change in the zero-field re- figurations. sistance, R(0), is dominated by the sharp temperature dependence of the thinner parts of the sample. The low- field features in magnetoresistance, however, are mainly dominant strongly indicates a proximity effect between contributed by the thicker parts, where the dephasing the IF and the TI. In figs. 2 & 3 we presented the nega- length is known to be longer.11 Therefore, after dividing tive MR in the same sample, whereas four bilayer samples by R(0) as the normalization factor, ∆R(H)/R(0) ap- (labeled as BL1–4) with t . 4QL from different growth pears with a reduced magnitude. On the other hand, batches all demonstrated such proximity effect in a con- measuring the Hall effect, which is insensitive to film sistent manner. The thickness criterion (t . 4QL) co- inhomogeneity,29 the sheet carrier density is calculated 13 −2 incides with the thickness when the two surfaces of a showing similar values (n2D ≈ 2 × 10 cm ) for both 9 Bi2Se3 film are observed to be coupled, suggesting a TI-only films and TI-IF bilayers and are very close to surface-originated WL mechanism. reported values for MBE films.11 Below 2K (fig. 3d), we observed a continuous sharpen- So far in our discussions, we assumed that the low-field ing of the low-field WL feature when lowering the tem- negative magnetoresistance has an orbital origin such as perature, as expected from diminishing thermal dephas- WL effects suggested by recent theories.16–18 The pres- ing. Unexpectedly, the magnitude of negative MR was ence of a magnetic material brings possibilities of MR reduced when lowering the temperature below 1K. This originated from scatterings at localized spins.30 How- can be explained by the inhomogeneity observed in the ever, the magnitude of any spin-introduced MR should Bi2Se3 layer grown on top of EuS. While the Bi2Se3 thin be greater above TC and should diminish rapidly be- films grown on bare substrates were verified by AFM low TC as the spins being aligned during the ferromag- and XRD to be adequately uniform in thickness, the netic transition,31 which is contrary to what we observed. TEM images taken at different locations on the cross- Moreover, we show in fig. 4a the broadening of he nega- section of the BL3 sample show large variations (±2QL) tive MR as the sample is rotated from perpendicular to in thickness of the Bi2Se3 layer, with an estimated mean near parallel applied magnetic field, clearly indicating its value consistent with the thickness of the Bi2Se3-only film orbital origin. The fact that some negative MR features grown in the same PLD session. For the thicknesses of in- remain finite in the near-parallel field is likely due to the terest (t . 4QL), both the resistance of Bi2Se3 films and uneven thickness in the Bi2Se3 layer as a consequence of a its temperature dependence are known to change sharply locally slanted top surface, on which local electron trans- with thickness at low temperatures.11 Thus special diffi- port may have a finite angle to the magnetic field that culty is introduced when measuring the sheet resistance is parallel to the sample plane. In fact, when the mag- with a van der Pauw method, where the sample thick- netic field is scaled with an effective angle θ∗ between the ness is assumed to be uniform.27,28 With such inhomo- magnetic flux and the normal of local electron transport, geneous geometry, electric conduction is limited by the both the WL at low fields and the WAL at higher fields 4 are found to coincide well with the perpendicular-field that a recent study of Bi2Se3/EuS bilayers shows a weak MR. The difference between the effective angles (θ∗) and tendency towards the anomalous Hall effect, which was the nominal angles read from the instrument (θ) is small argued to indicate emergence of a ferromagnetic phase for most angles (|θ − θ∗|≤ 3◦ for θ ≤ 60◦) and increasing in TI.38 Samples in that study consisted of a bottom ∗ ◦ ◦ towards parallel field (θ ≈ 75 for θ = 89 ). This is Bi2Se3 film and a top EuS film, making it impossible consistent with expectations from a film with a smooth to determine whether the EuS is truly insulating. We bottom surface and an uneven top surface. contrast it with the present approach in which we first Theory predicts a gap opened at the Dirac point as obtained high-quality, well characterized EuS films, on a result of a proximity between the TI and the IF.32 which the TI film was deposited. Since a true 3D TI will Such gap-opening is expected to result in WL only when have one and only one surface state, irrespective of in- the Fermi level is sufficiently near the gap.16–18 However, homogeneities and local defects, we prefer this approach as is common in ungated and uncompensated Bi2Se3, and further argue that the unusual negative MR at low our carrier densities suggest a Fermi level intersecting fields below the Curie temperature of the ferromagnet is the conduction band and therefore far away from the a strong indication of a proximity effect between a topo- Dirac point.33 Recent calculations suggest that surface logical insulator and an insulating ferromagnet. charges on the IF layer may result in band-bending at the interface,34 hence providing a possible mechanism to move the gap towards the Fermi level. However, whether ACKNOWLEDGMENTS this is the case depends on the details of atomic arrange- ment at the interface,35 which cannot be determined with available data. On the other hand, low levels of sul- We thank Ion Garate, Xiaoliang Qi, Boris Spivak, fur doping are known to modify the band structure of Nicholas Breznay, Gerwin Hassink and Philip Wu for 36,37 Bi2Se3 while preserving its structural phase. Such helpful discussions. Qi is grateful for previous guid- an effect may as well be present at the Bi2Se3-EuS inter- ance of Yoshifumi Tokiwa, Philipp Gegenwart and Malte face. Attempts to perform ARPES measurements to de- Grosche. This work is supported by DARPA, MesoDy- termine the band gap, the location of the Dirac point and namic Architecture Program (MESO) through the con- the Fermi level produced inconclusive results, primarily tract number N66001-11-1-4105, by FENA, and by a seed due to the inhomogeneity in films thicknesses. We note grant from DOE for the study of TI.

∗ [email protected] and A. Kapitulnik, Appl. Phys. Lett. 101, 153105 (2012). 1 X.-L. Qi and S.-C. Zhang, Rev. Mod. Phys. 83, 1057 13 G. Bergmann, Phys. Rep. 107, 1 (1984). (2011). 14 L. Zhang, M. Dolev, Q. I. Yang, R. H. Hammond, B. Zhou, 2 M. Z. Hasan and C. L. Kane, Rev. Mod. Phys. 82, 3045 A. Palevski, Y. Chen, and A. Kapitulnik, “Weak localiza- (2010). tion effects as evidence for bulk quantization in thin films 3 J. E. Moore, Nature 464, 194 (2010). Bi2Se3,” (2013), arXiv:1307.5358. 4 X.-L. Qi, R. Li, J. Zang, and S.-C. Zhang, Science 323, 15 A. A. Taskin, S. Sasaki, K. Segawa, and Y. Ando, Phys. 1184 (2009). Rev. Lett. 109, 066803 (2012). 5 H. Zhang, C.-X. Liu, X.-L. Qi, X. Dai, Z. Fang, and S.-C. 16 H.-Z. Lu, J. Shi, and S.-Q. Shen, Phys. Rev. Lett. 107, Zhang, Nat. Phys. 5, 438 (2009). 076801 (2011). 6 Z. Alpichshev, R. R. Biswas, A. V. Balatsky, J. G. Ana- 17 I. Garate and L. Glazman, Phys. Rev. B 86, 035422 (2012). lytis, J.-H. Chu, I. R. Fisher, and A. Kapitulnik, Phys. 18 H.-Z. Lu and S.-Q. Shen, Phys. Rev. B 84, 125138 (2011). Rev. Lett. 108, 206402 (2012). 19 M. Lang, L. He, X. Kou, P. Upadhyaya, Y. Fan, H. Chu, 7 J. G. Checkelsky, Y. S. Hor, M.-H. Liu, D.-X. Qu, R. J. Y. Jiang, J. H. Bardarson, W. Jiang, E. S. Choi, Y. Wang, Cava, and N. P. Ong, Phys. Rev. Lett. 103, 246601 (2009). N.-C. Yeh, J. Moore, and K. L. Wang, Nano Lett. 13, 48 8 Y. Xia, D. Qian, D. Hsieh, L. Wray, A. Pal, H. Lin, A. Ban- (2013). sil, D. Grauer, Y. S. Hor, R. J. Cava, and M. Z. Hasan, 20 M. Liu, J. Zhang, C.-Z. Chang, Z. Zhang, X. Feng, K. Li, Nat. Phys. 5, 398 (2009). K. He, L.-l. Wang, X. Chen, X. Dai, Z. Fang, Q.-K. Xue, 9 Y. Zhang, K. He, C.-Z. Chang, C.-L. Song, L.-L. Wang, X. Ma, and Y. Wang, Phys. Rev. Lett. 108, 036805 (2012). X. Chen, J.-F. Jia, Z. Fang, X. Dai, W.-Y. Shan, S.-Q. 21 Q. I. Yang, J. Zhao, L. Zhang, M. Dolev, A. D. Fried, Shen, Q. Niu, X.-L. Qi, S.-C. Zhang, X.-C. Ma, and Q.-K. A. F. Marshall, S. H. Risbud, and A. Kapitulnik, “Pulsed Xue, Nat. Phys. 6, 584 (2010). laser deposition of high-quality thin films of the insulating 10 H.-T. He, G. Wang, T. Zhang, I.-K. Sou, G. K. L. Wong, ferromagnet eus,” (2013), arXiv:1308.3820. J.-N. Wang, H.-Z. Lu, S.-Q. Shen, and F.-C. Zhang, Phys. 22 J. Xia, P. T. Beyersdorf, M. M. Fejer, and A. Kapitulnik, Rev. Lett. 106, 166805 (2011). Applied Physics Letters 89, 062508 (2006). 11 Y. S. Kim, M. Brahlek, N. Bansal, E. Edrey, G. A. Kapile- 23 A. D. Fried, M. M. Fejer, and A. Kapitulnik, “A scanning, vich, K. Iida, M. Tanimura, Y. Horibe, S.-W. Cheong, and all-fiber sagnac interferometer for high resolution magneto- S. Oh, Phys. Rev. B 84, 073109 (2011). optic measurements at 820nm,” To be published. 12 L. Zhang, R. Hammond, M. Dolev, M. Liu, A. Palevski, 24 Q. I. Yang, M. Liu, L. Zhang, et al., “PLD growth of 5

topological insulators Bi2Se3 and Bi2Te3 thin ﬁlms,” To a preprint on GdN-Bi2Se3 heterostructure by A. Kandala, be published. A. Richardella, D. W. Rench, D. M. Zhang, T. C. Flana- 25 S. Hikami, A. I. Larkin, and Y. Nagaoka, Prog. Theor. gan and N. Samarth (http://arxiv.org/abs/1212.1225), Phys. 63, 707 (1980). where similar WL behavior was observed for Fermi levels 26 D. Khmel’nitskii, Physica B+C 126, 235 (1984). intersecting the conduction band. 27 L. Van der Pauw, Philips Tech. Rev. 20, 220 (1958). 34 W. Luo and X.-L. Qi, Phys. Rev. B 87, 085431 (2013). 28 R. Chwang, B. Smith, and C. Crowell, Solid-State Elec- 35 X.-L. Qi, Private communication. tronics 17, 1217 (1974). 36 C. Iona, E. Hatzikraniotis, K. M. Paraskevopoulos, and 29 H. J. Juretschke, R. Landauer, and J. A. Swanson, J. T. Kyratsi, AIP Conference Proceedings 1203, 997 (2010). Appl. Phys. 27, 838 (1956). 37 M. Neupane, S.-Y. Xu, L. A. Wray, A. Petersen, 30 T. Micklitz, T. A. Costi, and A. Rosch, Phys. Rev. B 75, R. Shankar, N. Alidoust, C. Liu, A. Fedorov, H. Ji, J. M. 054406 (2007). Allred, Y. S. Hor, T.-R. Chang, H.-T. Jeng, H. Lin, A. Ban- 31 C. Haas, Phys. Rev. 168, 531 (1968). sil, R. J. Cava, and M. Z. Hasan, Phys. Rev. B 85, 235406 32 R. Yu, W. Zhang, H.-J. Zhang, S.-C. Zhang, X. Dai, and (2012). Z. Fang, Science 329, 61 (2010). 38 P. Wei, F. Katmis, B. A. Assaf, H. Steinberg, P. Jarillo- 33 During the submission of this paper, we were informed of Herrero, D. Heiman, and J. S. Moodera, Phys. Rev. Lett. 110, 186807 (2013).