Table Of Content

Construction of a 3He magnetic force microscope with a vector magnet Jinho Yang,1,2 Ilkyu Yang,1 Yun Won Kim,1,a) Dongwoo Shin,1,2 Juyoung Jeong,1,2 Dirk Wulferding,1 Han Woong Yeom,1,2 and Jeehoon Kim1,2,b) 1)Center for Artificial Low Dimensional Electronic Systems, Institute for Basic Science, 77 Cheongam-Ro, Nam-Gu, Pohang 790-784, Korea 2)Department of Physics, Pohang University of Science and Technology, Pohang 790-784, Korea (Dated: 19 January 2016) Weconstructeda3Hemagneticforcemicroscopeoperatingatthebasetemperatureof300mKunderavector magneticfieldof2-2-9Tinthex−y−zdirection. Fiberopticinterferometryasadetectionschemeisemployed inwhichtwohome-builtfiberwalkersareusedforthealignmentbetweenthecantileverandtheopticalfiber. The noise level of the laser interferometer is close to its thermodynamic limit. The capabilities of the sub- 6 1 Kelvinandvectorfieldaredemonstratedbyimagingthecoexistenceofmagnetismandsuperconductivityina 0 ferromagneticsuperconductor(ErNi B C)atT=500mKandbyprobingadipoleshapeofasingleAbrikosov 2 2 2 vortex with an in-plane tip magnetization. n a J I. INTRODUCTION allows investigating phenomena emerging under extreme 8 conditions such as superconductivity in heavy fermions, 1 Atomicforcemicroscopy(AFM)isoneofthemostver- anisotropic superconductivity, and the interplay of mag- satile imaging techniques to investigate surface science netism and superconductivity. ] t at the nanoscale. Magnetic force microscopy (MFM),1 e a variant of AFM with a magnetized tip, plays an im- d portant role in studying the formation of magnetic do- - s mains in magnetic materials,2–4 vortex dynamics in su- II. INSTRUMENTATION n perconductors,5–12 and exotic spin structures such as i . skyrmions.13 So far, several home-built low-temperature A. 3He Cryostat insert & dewar with a vector magnet s c MFMs have been operated down to the 4He tempera- i ture within a single axis magnet.3,14–18 On the other s We employ a custom-built 3He cryostat and a low-loss y hand, MFM apparatus operating under high demanding dewar with a vector magnet from JANIS [see Fig. 1(a)]. h conditions, such as sub-Kelvin temperatures and vector The outer jacket of the dewar is evacuated to a pres- p magnetic fields, are rare,19,20 although they are useful [ sure lower than 10−5 Torr prior to cooldown. The liquid 1 fsourpeurncdoenrdsutacntidviintyg21unacnodnvqeunatniotunmal/(me.ogl.e,cuhleaarvmyafgernmetiiosnm) 4He consumption is about 10 liters a day without magnetic field applied. To minimize the effect of heat radia- v at the nanoscale under extreme conditions.22 At present tion, baffles are installed inside the displacer. The heat 8 the demand for a sub-Kelvin MFM is high for inves- loadfromelectricalwireshasbeenminimizedbywinding 1 tigating f-electron physics as a recent scanning tun- 4 them around thermal anchors which are attached to the neling microscopy study resolves the competition of f- 4 1-K pot. We control the sample temperature via char- and d-electrons in heavy fermion systems.23,24 In addi- 0 coal,1-Kpot,and3Hepot[seeFig. 1(a)]. Thebasetem- 1. tion, the use of a vector magnet allows further insight perature(T=300mK)ofthe3Hepotisreachedwithin4 intoanisotropyinsuperconductivitysuchasnon-uniform 0 hoursfromT=4Kandremainsfor8hoursifinitiallyall pairing symmetry and magnetic penetration depth.25 6 25 L of 3He gas were condensed. The three-axis vector 1 In this Article we report the construction of a 3He magnet is located at the bottom of the dewar as shown v: MFMwithavectorfieldof2-2-9Tforthex−y−z direc- inFig. 1(a)andimmersedinliquid4Hecryogen. Wecan tionandabasetemperatureofT=300mK.Wecalibrate i applythevectorfieldof2-2-9Tineachx−y−zdirection X the system’s stray field by imaging Abrikosov vortices as with a field homogeneity of ±0.1% over a sphere of 1 cm r a function of field, and demonstrate its low temperature diameter at the sample position. The electronics for the a and vector field capabilities by resolving the coexistence vector magnet are supplied by Cryomagnetics Inc., and of superconductivity and magnetism in a ferromagnetic controlled by a home-made LabVIEW-based program. superconductor at T= 500 mK, as well as by imaging the change of a single vortex shape depending on the direction of the tip magnetization. Our 3He MFM system B. MFM head a)PresentAddress: DivisionofPhysicalMetrology,KoreaResearch The MFM head consists of three parts: (i) the tip InstituteofStandardsandScience,Daejeon305-340,Korea stage,(ii)thesamplestage,and(iii)thescanner&x,y,z b)Electronicmail: [email protected] coarse positioners. 2 3He Reservoir (a) 500 mm 50 mm 20 mm 3He Pot Fiber winding Sample Stage Displacer spool Cantilever Dewar 24 pin connectors & Optical Fiber (4He Reservoir) Fiber Walkers Charcoal Head housing 1K Pot (macor) 3He Pot Vector Magnet Scanner & Heat shield X,Y,Z Walkers (gold-coated copper) 24 pin connectors MFM Head Part (b) 2 mm walking range Electrical Sample stage connections sliding type Heat shield Optical fiber (gold-coated copper) FIG.1. (a)Schematiccross-sectionforthefrontviewofthewholeMFMsystem,azoom-inontheMFMhead,andazoom-in on the microscope. (b) Top view of the sample stage. 1. Tip stage the z walker regulates the tip-fiber position at the maximum slope of the interference curve to achieve a maximum sensitivity. Considering a sinusoidal waveform for Thetipstageconsistsofthecantilever,anopticalfiber, the interference signal, I(D) = sin(2πD/λ), the maxi- and two fiber walkers, as shown in Fig. 1(a). It detaches mum slope of the signal is reached at d2I(D) = 0, where dD2 conveniently from the MFM head, allowing an easy re- D is the tip-fiber distance and λ is the wavelength (1550 placementofacantilever. Thecantileverissecuredviaa nm) of the laser. This condition results in: Be-Cuspringwhichalsoestablishesanelectricalconnec- tionforelectricscanningprobemicroscopymodessuchas D =(2n+1)λ/8, (1) conducting AFM, electric force microscopy, and Kelvin force microscopy. We use an optical interferometric sys- wherenisanintegernumber. Atthispoint,thesensitiv- tem for measuring the cantilever motion that is directly ity of the MFM signal (I) reaches its maximum, as dI dD proportional to the force between the tip and sample. A is maximized.27 mechanism to optimize the alignment between the cantilever and the fiber end is of importance for low temperature applications.26 Therefore, we adopt two com- 2. Sample stage pact home-built walkers to control the relative position of the fiber and the tip in the x-z direction: x for the The sample stage is attached directly to the 3He pot cantilever’s center in width and z for the fiber-tip dis- via four Cu rods, allowing a good thermal anchoring. tance. The intensity and sensitivity of the MFM signal A horizontal sliding mechanism for replacing the sample can be maximized by the following procedure. First, the holder ensures a safe and easy mounting procedure. The x walker locates the fiber end at the middle of the can- Au-plated Cu sample holder has a resistive heater ele- tilever to achieve the maximum signal intensity. Then, ment(a9-Ωchipresistorforsurfacemount)andacernox 3 50 (a) +0.6 (b) RS: Reference signal IS: Interference signal optical box λ = 1550 nm FILTER LASER z (nm) Phase (deg) laser in IS 100% MFCMANT hILEeVEaRd 0 2 µm -2.2 1 µm DEPTHEOCTTOOR RISS RS 100I%S 90% C O90U P: 1L0ER RS 9R0S% 10% FIG.2. (a)Topographicimageofthecalibrationsamplefrom ATTENUATOR MikroMasch. (b)MFMsignalfromadisassembledharddisk drive. FIG. 3. Schematics of the optical interferometric system. temperaturesensorforanadditionaltemperaturecontrol together with the 3He-pot control. We can study several samples during a single cooldown with our multi-sample width) for the scanner calibration. A resulting topo- stage [see Fig. 1(b)] due to the large walking range of graphic image obtained at room temperature is shown the piezo walkers. This not only avoids lengthy warm- in Fig. 2(a). In Fig. 2(b) we show the MFM image of a up and cool-down procedures for changing samples, but hard disk drive obtained in amplitude-modulation mode it also allows to perform comparative studies among the at ambient conditions with the tip-lift height of 100 nm. installedsamplesundercontrolledconditions,confirming Individual magnetic bits are clearly resolved. that the geometry of the cantilever, the tip magnetization, etc. remain unchanged between the measurements. The sample holder is grounded electrically to the MFM C. Optical interferometric system head [see Fig. 1(a)]. The optical interferometer system provides a stable, low-noise environment for the detection of a cantilever 3. Scanner & walkers motion.26–29 We use a 2×2 coupler with an intensity ratio of 9:1 (Gould Fiber Optics). A solid state laser Ingeneral,apiezotubescanner,routinelyusedinscan- (S3FC1550, THORLABS) of λ=1550 nm offers a good ning tunneling microscope, offers a simple, cost-efficient, controloftemperatureandpower. Thephotonenergyof and rigid scanning solution in spite of a crosstalk be- thelaser(λ=1550nm=0.8eV)issmallerthantheband (cid:98) tweenz andx−y motionaswellasasmallscanrangein gapofaSicantilever(E =1.1eV),andthustransfersless g the z direction. For magnetic force microscopic applica- heatthanalaserofsmallerwavelength. Weminimizethe tionsthelargescanareaisrequiredtoimagetheµm-size laser fluctuation by adjusting the laser power and tem- magneticdomains. Wethereforeemployacompact,com- peraturebymonitoringthestabilityoftheinterferogram. mercially available Attocube scanner (ANSxyz100). Its Theadjustedlaserpowerdeliveredtothecantileverisas principleisbasedonapiezo-drivenleverarmmechanism small as 40 µW to avoid a direct cantilever heating. The and each axis is decoupled from one another. It offers a laser power and temperature on operation are typically scan range of 50×50×24 µm3 at 300 K and 30×30×15 0.34mWand23◦C,respectively. Notethatsuchoptimal µm3 at4K,respectively. WeusethreeAttocubewalkers parameters may vary even with the same laser model. (Attocube, ANPx101/RES, ANPz/RES) with a travel We use angled physical contact (APC) plugs for all op- distance of 5 mm each for a coarse positioning. An elec- tical connections to minimize undesired reflections from tricalconnectiontothescannerandwalkersisestablished optical connectors though the laser and the photo de- through a home-built 24-pin connector at the bottom of tector have built-in physical contact (PC) plugs. There- the head [see Fig. 1(a)]. Manganin wires with low ther- fore, we use a custom spec adaptor to connect PC with malandelectricalconductivityandCuwiresareusedfor APC. The schematic diagram for the interferometer sys- the scanner and the piezo walkers, respectively, because temisdepictedinFig. 3. Theincidentlaserbeampasses the high resistance of manganin wires results in a large throughanopticalisolator(ThorlabsIO-H-1550APC)to decay time constant, which is detrimental for the perfor- discriminate any side bands, and is split into two beams mance of a stick-slip walker. The decay time constant via an optical coupler. 90% of the incident beam goes of a driving voltage should be as small as 10 µsec for into the photo detector (new focus model 2117) through a stick-slip walker: For example, the capacitance of the an attenuator and serves as a reference signal (blue dot- Attocube walker is 5 µF, requiring the wire resistance is ted lines in Fig. 3). The remaining 10% are fed into the as small as 2 Ω. microscopehead,wheretheypartlytransmitthroughthe We use a standard sample (TGXYZ01/NM from cleaved fiber end, and reflect at the cantilever. The re- MikroMasch with 20 nm step height and 3 µm step maining part of the beam is reflected at the fiber end, 4 wooden table Pohang, as shown in Fig. 4. A 20-ton concrete block (lead-filled) resting on the heavy duty damping springs serves as the dewar with pneumatic base for our system. Three pneumatic legs standing on 3He-MFM insert legs the concrete block carry a home-built triangular wooden concrete block table filled with lead bricks of 800 kg to further lower the system’s resonance frequency. The 3He cryostat insert is fixed on the table, and the dewar with a vector magnet moves up anddownfor experiments throughthe rectangular pit. We characterize the thermal noise level of our micro- damping springs scope as follows:30 (cid:115) f k TB δfmin = 0 B , (3) 2πkQA2 wheref ,k ,T,B,k,Q,andAareresonancefrequency, 0 B Boltzmann constant, temperature, measurement band- width, spring constant, quality factor, and oscillation amplitude, respectively. Based on our typical operation parameters (f = 75195 Hz, B = 100 Hz, k = 2.8 N/m, 0 Q=48554, and A=4.8 nm) at T=4.2 K, the minimum FIG. 4. Schematics of the vibration isolation laboratory. detectable frequency shift is δfmin = 4.6 mHz, and the corresponding minimal detectable force gradient is: andinterfereswiththereflectedlightfromthecantilever backside within the fiber; this interference signal is fed (cid:12)(cid:12)(cid:12)∂Fzmin(cid:12)(cid:12)(cid:12)=2k· δfmin =3.4·10−7N/m. (4) into the coupler (see Fig. 3). Both the reference and the (cid:12) ∂z (cid:12) f0 interferencesignalarefedintothebalancedphotodetec- The measured background, as shown by the red pro- tor,whichallowstheDCpartoftheinterferencesignalto file line in Fig. 5(a), varies by about 7.4 · 10−7 N/m, bezerosinceitissubtractedfromtheinterference. Thus, which only deviates by a factor of 2 from the thermody- the removal of the DC signal results in a wide dynamic namic noise limit. Additional noise fluctuations can be span of an amplifier allowing an increase of the DC gain attributed to electrical noise, mechanical vibrations, and as well as the subtraction of DC laser fluctuations. The possible magnetic inhomogeneities at the measurement interference intensity measured by a photo detector is as position. The MFM signal magnitude of an Abrikosov follows: vortex in the Nb film [black profile line in Fig. 5(a)] ex- (cid:18) (cid:19) 4πD ceeds the noise level by a factor of about 40, which is I =I −I cos , (2) 1 2 λ comparable to the previous report.18 where I is the laser power and I is proportional to I 1 2 1 multiplied by the optical visibility of the system, D is B. MFM results the fiber-cantilever distance, and λ is the wavelength of the laser.18 For a fixed wavelength, the intensity change Here we present two experimental results demonstrat- due to the cantilever motion is the most sensitive when ing the performance of our home-built 3He MFM. The 4πD/λ is close to a multiple of π/2, because the vari- images of Abrikosov vortices obtained in a Nb film high- ation in intensity due to the change of D is maximum lightthevectormagnetcapability. Thesimultaneousim- at these points: For example, dI is maximized when age of ferromagnetism and superconductivity at T=500 dD 4πD/λ is a multiple of π/2. Therefore we regulate the mK in ErNi B C proves the sub-Kelvin operation. 2 2 fiber-cantilever distance at these crossing points by the fiber piezo. 1. Vector magnet experiments III. CHARACTERIZATION AND PERFORMANCE The images of vortices in Fig. 5, obtained from a Nb film, show the vector magnet capability of our MFM. A. Vibration isolation and noise characterization The Nb thin film was grown on a Si substrate via electron beam deposition with a film thickness and critical In order to minimize the mechanical noise, our micro- temperature of 300 ± 5 nm and T = 8.8 K, respec- c scopewillbeinstalledintoanewlyconstructedvibration tively. The tip was magnetized along the z direction, isolation laboratory at the Institute for Basic Science in perpendicular to the surface of the Nb film, by applying 5 +30 (a) +38 (b) 20 (a) T = 0.5 K (d) Background By Bx Vortex 1 µm 1 µm 0 -38 5 µm 0 max (b) T = 1.0 K (c) T = 5.0 K (c) (d) |B | ~ 6.7 G stray 5 µm 5 µm -5 G +4 G FIG. 6. (a) MFM image of the magnetic superconductor ErNi B C, obtained at T = 500 mK. (b) and (c): Subse- 2 2 +6 G +10 G quent images obtained on the same scan area at 1.0 K and min 5.0 K, respectively. All images were taken with a tip-sample distance of 700 nm. (d) Comparison of Meissner force curves from ErNi B C at 500 mK (dashed black curve) and a refer- 2 2 FIG. 5. (a) Abrikosov vortices in a superconducting Nb film ence sample (Nb, solid red curve) to determine λ .31 L imagedwithanout-of-planemagnetizedMFMtip. Linepro- files of the vortex and the background are presented as black and red solid lines. (b) Vortices imaged at the same sample sign with field, as indicated by the MFM contrast. The position with an in-plane magnetized MFM tip. The insets number of vortices per scan frame as a function of fields in (a) and (b) indicate the magnetic field direction of the tip is shown in Fig. 5(d). The experimental flux quantum andvortex. (c)FourrepresentativeMFMimagesobtainedat differentmagneticfields. (d)Numberofvorticesona10×10 Φexp can be obtained from Fig. 5(d) as follows: n = µm2 area as a function of the out-of-plane applied magnetic Φex1pAB +Bstray, where n is a number of vortices for a field. All images were taken at T = 4.2 K and with a tip- scan area of A, B is the stray field. We find a stray stray sample distance of 300 nm. field of B = +6.7 G, and for a given area of A=100 stray µm2, we obtain Φ = 20.02 ± 0.84 Gµm2, which is exp in good agreement with the theoretical quantum flux of a magnetic field of 10 kOe larger than the tip coercive Φ0 =h/2e≈20.67 Gµm2. field (∼ 300 Oe). We field-cool the sample from T = 12 K (above T ) in the presence of several Oersted. Due to c local pinning centers the penetrating flux quanta can be 2. Ultra-low temperature experiments trappedwithinthescanframe,whilethesampleremains in the pure Meissner state. The resulting MFM image In order to demonstrate the low temperature capa- is shown in Fig. 5(a): the inset sketches the field lines bility, we investigate magnetism in ErNi B C, a mag- 2 2 from a single vortex. The vortex image with an in-plane netic superconductor, at the sub-Kelvin temperature.31 tip magnetization is shown in Fig. 5(b) after polarizing It turns superconducting below T = 10.5 K and de- c the tip magnetization along the in-plane direction by a velops a weak ferromagnetic order below T = 2.3 WFM vector magnet. The scan areas in Fig. 5(a) and (b) are K within the superconducting phase. The MFM im- identical. TheMFMsignalofavortexresemblesadipole, age shown in Fig. 6(a) was obtained at T = 500 mK, which is in contrast to the monopole signal of an out-of- and clearly shows magnetic stripes as well as Abrikosov plane tip magnetization directions [see Fig. 5(a)]. The vortices (dark spots), indicating a coexistence of mag- two different vortex images result from the two differ- netism and superconductivity. At temperatures higher ent tip magnetizations. In general, the force gradient is than T the bright stripes vanish as shown in Figs. WFM positiveforanattractiveinteractionbetweenthetipand 6(b) and (c), signaling the fading of the magnetic order vortex as it is negative for a repulsive interaction. The upon reaching the ferromagnetic transition. At elevated tip magnetized with an in-plane field experiences both temperaturesthevorticestendtobemobilebyovercom- positive and negative force gradients as it crosses a sin- inglocalpinningpotentials,whichallowsamanipulation gle vortex, which results in dipole-like vortex as shown of vortices via the tip magnetic moment upon scanning. in Fig. 5(b). This result displays a good demonstration This is evident from the smeared-out vortex lines in Fig. of a vector field effect. 6(c). The absolute values of the local London penetra- Imaging vortices as a function of applied field allows tion depth (λ ) can be obtained through a comparative L us to estimate the stray field of our setup. The vortex Meissner technique, as shown in Fig. 6(d).32 When the imagesatfourdifferentmagneticfieldsareshowninFig. sample is in the superconducting state, the magnetic tip 5(c). Note that the polarity of the vortices changes the experiences a Meissner force which grows algebraically 6 with decreasing the tip-sample distance. Using a com- 6A.Volodin,K.Temst,C.VanHaesendonck,Y.Bruynseraede,M. parative method, one can extract the absolute value of I.Montero,andI.K.Schuller,Europhys.Lett.58,582(2002). λ by overlaying the Meissner force curve of the sample 7A. Schwarz, U. H. Pi, M. Liebmann, R. Wiesendanger, Z. G. L Khim,andD.H.Kim,Appl.Phys.Lett.88,012507(2006). (dashed black curve) with that of a Nb reference sample 8M.RosemanandP.Gru¨tter,J.Appl.Phys.91,8840(2002). (solid red curve), λL,Nb = 110 nm. The addition of the 9U. H. Pi, A. Schwarz, M. Liebmann, R. Wiesendanger, Z. G. shifteddistanceof100nmtoλL,Nb leadstoapenetration Khim,andD.H.Kim,Phys.Rev.B73,144505(2006). depth in ErNi B C of λ =210 nm. This straight- 10L.Luan,O.M.Auslaender,D.A.Bonn,R.Liang,W.N.Hardy, 2 2 L,ENBC forward method to measure λ , one of the two intrinsic andK.A.Moler,Phys.Rev.B79,214530(2009). L superconductingparameters,atsub-Kelvintemperatures 11E. W. J. Straver, J. E. Hoffman, O. M. Auslaender, D. Rugar, andK.A.Moler,Appl.Phys.Lett.93,172514(2008). under a vector field allows further investigation of exotic 12I. Yang, J. Yang, D. Wulferding, S. Chung, H. Yeom, K. Kim, superconductivity together with anisotropy. andJ.Kim(unpublished). 13P.Milde,D.Köhler,J.Seidel,L.M.Eng,A.Bauer,A.Chacon, J. Kindervater, S. Mu¨hlbauer, C. Pfleiderer, S. Buhrandt, C. IV. SUMMARY Schu¨tte,andA.Rosch,Science340,1076(2013). 14H. J. Hug, Th. Jung, and H.-J. Gu¨ntherodt, Rev. Sci. Instrum. 63,3900(1992). Insummary,weconstructeda3HeMFMoperatingata 15R.Euler,U.Memmert,andU.Hartmann,Rev.Sci.Instrum.68, basetemperatureof300mKundervectormagneticfields 1776(1997). 16M.RosemanandP.Gru¨tter,Rev.Sci.Instrum.71,3782(2000). of H =2-2-9 T. We employed a low-noise fiber inter- xyz 17T.-M.ChuangandA.deLozanne,Rev.Sci.Instrum.78,053710 ferometer system with a noise level close to its thermo- (2007). dynamic limit. The sub-Kelvin and vector field capabili- 18E. Nazaretski, K. S. Graham, J. D. Thompson, J. A. Wright, ties of the apparatus were demonstrated by imaging the D. V. Pelekhov, P. C. Hammel, and R. Movshovich, Rev. Sci. shape change of the Abrikosov vortices from monopole Instrum.80,083704(2009). 19D.V.Pelekhov,J.B.Becker, andG.Nunes, Rev.Sci.Instrum. to dipole according to the rotation of the tip magneti- 70,114(1999). zation from out-of-plane to in-plane in a Nb film and 20O¨. Karci, M. Dede, and A. Oral, Rev. Sci. Instrum. 85, 103705 by resolving a coexistence of magnetism and supercon- (2014). ductivity in ErNi B C at T=500 mK. Our 3He MFM 21P. Thalmeier, G. Zwicknagl, G. Sparn and F. Steglich, “Super- 2 2 opensthedoortoinvestigatemagnetisminexoticsuper- conductivity in Heavy Fermion Compounds”, in “Frontiers in Superconducting Materials”, ed. A. V. Narlikar, (Springer Ver- conducting phenomena emerging at sub-Kelvin temper- lag2005),p.109. atures, such as magnetically mediated superconductiv- 22S.Sachdev,NaturePhysics4,173(2008). ity in heavy fermion systems and the nonlinear Meissner 23P. Aynajian, E. H. da Silva Neto, A. Gyenis, R. E. Baumbach, effect with an anisotropic pairing symmetry. It further- J. D. Thompson, Z. Fisk, E. D. Bauer, and A. Yazdani, Nature moreallowsustostudytheanisotropyofunconventional 486,201(2012). 24M. P. Allan, F. Massee, D. K. Morr, J. Van Dyke, A. W. Rost, magnetic domains in a vector magnetic field. A.P.Mackenzie,C.Petrovic,andJ.C.Davis,NaturePhysics9, 468(2013). 25D.Xu,S.K.Yip,andJ.A.Sauls,Phys.Rev.B51,16233(1995). ACKNOWLEDGMENTS 26A.Oral,R.A.Grimble,H.O¨.O¨zer,andJ.B.Pethica,Rev.Sci. Instrum.74,3656(2003). 27D. Rugar, H. J. Mamin, R. Erlandsson, J. E. Stern, and B. D. This work was supported by the Institute for Basic Terris,Rev.Sci.Instrum.59,2337(1988). Science (IBS), Grant No. IBS-R014-D1. 28D.D.Smith,J.A.Marohn,andL.E.Harrell,Rev.Sci.Instrum. 72,2080(2001). 1Y.MartinandH.K.Wickramasinghe,Appl.Phys.Lett.50,1455 29K. J. Bruland, J. L. Garbini, W. M. Dougherty, S. H. Chao, S. (1987). E.Jensen,andJ.A.Sidles,Rev.Sci.Instrum.70,3542(1999). 2J.Jeong,I.Yang,J.Yang,O.E.Ayala-Valenzuela,D.Wulferd- 30T. R. Albrecht, P. Gru¨tter, D. Horne, and D. Rugar, J. Appl. ing,J.-S.Zhou,J.B.Goodenough,A.deLozanne,J.F.Mitchell, Phys.69,668(1991). N.Leon,R.Movshovich,Y.H.Jeong,H.W.Yeom,andJ.Kim, 31D.Wulferding,I.Yang,J.Yang,M.Lee,H.C.Choi,S.L.Bud’ko, Phys.Rev.B92,054426(2015). P.C.Canfield,H.W.Yeom,andJ.Kim,Phys.Rev.B92,014517 3M.Liebmann,A.Schwarz,S.M.Langkat,andR.Wiesendanger, (2015). Rev.Sci.Instrum.73,3508(2002). 32J.Kim,L.Civale,E.Nazaretski,N.Haberkorn,F.Ronning,A. 4A. Schwarz, M. Liebmann, U. Kaiser, R. Wiesendanger, T. W. S. Sefat, T. Tajima, B. H. Moeckly, J. D. Thompson, and R. Noh,andD.W.Kim,Phys.Rev.Lett.92,077206(2004). Movshovich,Supercond.Sci.Technol.25,112001(2012). 5A. Moser, H. J. Hug, I. Parashikov, B. Stiefel, O. Fritz, H. Thomas,A.Baratoff,H.-J.Gu¨ntherodt,andP.Chaudhari,Phys. Rev.Lett.74,1847(1995).