Arxiv:1609.06182V4 [Cond-Mat.Supr-Con] 18 May 2018 C

NanoSQUIDs: Basics & recent advances

M. J. Mart´ınez-Pérez1 and D. Koelle1 1Physikalisches Institut – Experimentalphysik II and Center for Quantum Science in LISA+, UniversitätTübingen,Auf der Morgenstelle 14, D-72076 Tübingen,Germany (Dated: May 21, 2018) Superconducting Quantum Interference Devices (SQUIDs) are one of the most popular devices in superconducting electronics. They combine the Josephson effect with the quantization of magnetic flux in superconductors. This gives rise to one of the most beautiful manifestations of macroscopic quantum coherence in the solid state. In addition, SQUIDs are extremely sensitive sensors allowing to transduce magnetic flux into measurable electric signals. As a consequence, any physical observable that can be converted into magnetic flux, e.g., current, magnetization, magnetic field or position, becomes easily accessible to SQUID sensors. In the late 1980’s it became clear that downsizing the dimensions of SQUIDs to the nanometric scale would encompass an enormous increase of their sensitivity to localized tiny magnetic signals. Indeed, nanoSQUIDs opened the way to the investigation of, e.g., individual magnetic nanoparticles or surface magnetic states with unprece- dented sensitivities. The purpose of this review is to present a detailed survey of microscopic and nanoscopic SQUID sensors. We will start by discussing the principle of operation of SQUIDs, plac- ing the emphasis on their application as ultrasensitive detectors for small localized magnetic signals. We will continue by reviewing a number of existing devices based on different kinds of Josephson junctions and materials, focusing on their advantages and drawbacks. The last sections are left for applications of nanoSQUIDs in the fields of scanning SQUID microscopy and magnetic particle characterization, putting special stress on the investigation of individual magnetic nanoparticles.

CONTENTS C. Susceptibility measurements 16

I. Introduction 1 V. nanoSQUIDs for scanning SQUID microscopy 18 A. SQUID microscopes using devices on planar II. SQUIDs: some basic considerations 2 substrates 18 A. Resistively and capacitively shunted junction B. SQUID-on-tip (SOT) microscope 19 model 2 B. dc SQUID basics 3 VI. Summary and outlook 20 C. SQUID readout 4 1. Flux-locked loop 4 Acknowledgments 20 2. Voltage readout 4 3. Critical current readout & threshold References 20 detection 5 4. Dispersive read out 5 I. INTRODUCTION III. nanoSQUIDs: design, fabrication & performance 5 The superconducting quantum interference device A. nanoSQUIDs: design considerations 6 (SQUID) consists of a superconducting ring intersected B. nanoSQUIDs based on metallic by one (rf SQUID) or two (dc SQUID) Josephson junc- superconductors 7 tions. SQUIDs constitute, still at present, the most sen- 1. Sandwich-type SIS junctions 7 sitive sensors for magnetic ﬂux in the solid state1,2. For 2. Sandwich-type SNS junctions 8 more than 50 years, a plethora of devices exploiting this 3. Constriction junctions 9 property have been envisioned, fabricated and used in 4. Proximized structures 11 many ﬁelds of applications3. These devices include volt-

arXiv:1609.06182v4 [cond-mat.supr-con] 18 May 2018 C. NanoSQUIDs based on cuprate meters, current amplifiers, metrology standards, motion superconductors 12 sensors and magnetometers. One of the key applications of SQUIDs is in magnetometry. Here, a superconduct- IV. nanoSQUIDs for magnetic particle detection 13 ing input circuit (flux transformer) picks up the mag- A. Nanoparticle positioning 13 netic flux density B, captured by superconducting pick- 1. In-situ nanoparticle growth 14 up loops of some mm2 or cm2 area, and the induced cur- 2. Scanning probe-based techniques 14 rent is then (typically inductively) coupled to a SQUID. 3. Techniques based on chemical The figure of√ merit of√ SQUID magnetometers is the field functionalization 15 resolution SB = √SΦ/Aeff , which can reach values B. Magnetization measurements 15 down to about 1fT/ Hz. Here, SΦ is the spectral den- 2 sity of flux noise of the SQUID and Aeff is the effective making the interpretation of the results much more di- area of the magnetometer. rect and simple. While their sensitivity deteriorates To ensure good coupling from an input circuit to a rapidly when Hall sensors are reduced to the submicron SQUID, typically thin film multiturn input coils are in- size, miniaturized SQUID-based sensors can theoretically tegrated on top of a washer-type SQUID loop. Typical reach quantum limited resolution. thin film washer SQUIDs have lateral outer dimensions In this review, we give an overview on some basics of 19 of several 100 µm, the inner hole size is several tens of nanoSQUIDs and recent advances in the field. After a µm and the lateral size of the Josephson junctions is sev- brief description of some SQUID basics in section II, we eral µm. Such devices are fabricated by conventional thin will review in section III important design considerations film technology, including micropattering by photolithog- for optimizing nanoSQUID performance and the state of raphy. With the development of a mature junction tech- the art in fabrication and performance of nanoSQUIDs nology, based on sandwich-type Nb/Al-AlOx/Nb junc- based on low-Tc and high-Tc superconductors, with em- tions in the 1980s4, Nb based dc SQUIDs became the phasis on the various types of Josephson junctions used. most commonly used type of devices for various appli- Subsequently, we will review important applications of cations. At the same time, first attempts were started nanoSQUIDs, divided into two sections: Section IV gives to further miniaturize the lateral dimensions of SQUIDs, an overview on applications of nanoSQUIDs for magnetic including the Josephson junctions5. This was made pos- particle detection, and section V addresses nanoSQUIDs sible by advances in nanolithography6 and was motivated for scanning SQUID microscopy. We will conclude with by the development of the theory for thermal noise in a short section VI, which gives a summary and outlook. the dc SQUID7, which showed that the energy resolution ε = SΦ/(2L) of dc SQUIDs can be improved by reducing the SQUID loop inductance L and junction capacitance II. SQUIDS: SOME BASIC CONSIDERATIONS C, to eventually reach and explore quantum limited res- 8 olution of such devices . These developments have trig- The working principle of a SQUID is based on two gered the realization of miniaturized dc SQUIDs for the fundamental phenomena in superconductors, the flux- investigating of small magnetic particles and for imag- oid quantization and the Josephson effect. The fluxoid ing of magnetic field distributions by scanning SQUID quantization arises from the quantum nature of super- microscopy to combine high sensitivity to magnetic flux conductivity, as the macroscopic wave function describ- 9 with high spatial resolution. In 1984, Ketchen et al. ing the whole ensemble of Cooper pairs shall not in- presented the first SQUID microsusceptometer devoted terfere destructively. This leads to the quantization of to detect the tiny signal produced by micron-sized mag- the magnetic flux Φ threading a superconducting loop20, netic objects, and in 1983 Rogers and Bermon developed in units of the magnetic flux quantum Φ0 = h/2e ≈ the first system to produce 2-dimensional scans of mag- 2.07 × 10−15 Vs. 10 netic flux structures in superconductors . Both devel- The Josephson effect21,22 results from the overlap of opments were pushed further in the 1990s. Wernsdorfer 11,12 the macroscopic wave functions between two supercon- et al. used micron-sized SQUIDs to perform experi- ducting electrodes at a weak link forming the Josephson ments on the magnetization reversal of nanometric par- junction (JJ). The supercurrent Is through the weak link ticles, which were placed directly on top of the SQUIDs. and the voltage drop U across it satisfy the Josephson At the same time, scanning SQUID microscopes with relations miniaturized SQUIDs and/or pickup loop structures have been developed, at that time with focus on studies on Φ I (t) = I sin δ(t) (a) U(t) = 0 δ˙ (b) , (1) the pairing symmetry in the high transition temperature s 0 2π 13 (high-Tc) cuprate superconductors . Since then much effort has been dedicated to the further miniaturization with the gauge-invariant phase difference δ between the of SQUID devices and to the optimization of their noise macroscopic wave functions of both superconductors and 14 characteristics . the maximum attainable supercurrent I0; the dot refers Studies on the properties of small spin systems, such to the time derivative. The simple sinusoidal current- as magnetic nanoparticles (MNPs) and single molecule phase relation (CPR), Eq. (1(a)), is found for many kinds of JJs. However, some JJ types exhibit a non-sinusoidal magnets (SMMs), have fueled the development of new 23 magnetic sensors for single particle detection and imag- CPR, which can even be multivalued . ing with improved performance. Many of the recent advances in this field include the development of magneto- optical techniques based on nitrogen vacancy centers in A. Resistively and capacitively shunted junction diamond15,16 or the use of carbon nanotubes (CNTs) as model spin detectors17. Alternatively, miniature magnetometers, based on either microHall bars18 or micro- and A very useful approach to describe the phase dynamics nanoSQUIDs, provide direct measurement of the stray of a JJ is the resistively and capacitively shunted junction magnetic fields generated by the particle under study, (RCSJ) model24–26. Within this model, the current flow 3

(a) I B. dc SQUID basics

The dc SQUID27 is a superconducting loop (with in- Id Iqp Is ductance L) intersected by two JJs [Fig. 2(a)]. With an externally applied magnetic flux Φ through the loop, the C R I I U 0 N fluxoid quantization links the phase differences δ1 and δ2 of the two JJs to the total flux in the SQUID ΦT = Φ+LJ via 2π δ1 − δ2 + 2πn = (Φ + LJ) . (5) (b) i= Φ0 2EJ 0 Here, J is the current circulating in the SQUID loop and UJ n is an integer28. Defining the screening parameter as

2LI0 1/2 βL ≡ , (6) Φ0

1 one finds in the limit βL 1 a negligible contribution 0 1 2 of LJ to ΦT in Eq. (5), and by assuming for simplic- δ/2π ity identical values for I0 in the two JJs, the maximum supercurrent (critical current) I of the SQUID can be Fig. 1 c FIG. 1. RCSJ model: (a) Equivalent circuit. (b) Tilted wash- easily obtained as board potential for different normalized bias currents i. πΦ Ic = 2I0 cos . (7) is split into three parallel channels [Fig. 1(a)]: (i) a su- Φ0 percurrent Is [Eq. (1(a))], (ii) a dissipative quasiparticle current I = U/R across an ohmic resistor R and (iii) a The pronounced Ic(Φ) dependence [Fig. 2(b) for βL 1] qp can be used to probe tiny changes in applied magnetic displacement current Id = C ∂U/∂t across the junction capacitance C. A finite temperature T is included as a flux. No analytical expression for Ic(Φ) can be obtained thermal current noise source I from the resistor. With when a finite βL and hence a finite L is included, unless N restrictions are imposed to some of the important SQUID Kirchhoff’s law and Eq. (1(b)), one obtains the equation 26,29 of motion for the phase difference δ parameters . An increasing βL leads to a monotonic decrease of the critical current modulation ∆Ic/2I0 with Φ0 ˙ Φ0C ¨ I + IN = I0 sin δ + δ + δ . (2) 2πR 2π 2 β This is equivalent to the equation of motion of a point-like (b) L particle moving in a tilted washboard potential [Fig. 1(b)] (a) I /I 5 I c 0 UJ = EJ(1 − cos δ) − (i + iN)δ , (3) 1 1 with normalized currents i = I/I0, iN = IN/I0 and the J Josephson coupling energy E = I Φ /(2π). In this anal- J 0 0 0.01 ogy, the mass, friction coefficient, driving force (tilting 0 δ Φ δ 0 0.5 1.0 1.5 2.0 the potential) and velocity correspond to C, 1/R, I and 1 2 ΦΦ / 0 U, respectively. Hysteresis in the current voltage characteristics (IVC), i.e. bias current I vs time averaged L voltage V = hUi, can be understood as a consequence of (c) 100 ˙ ∝ the particle’s inertia: the dissipative state hδi ∝ V 6= 0 is ∆I /2I 1/βL achieved once the metastable minima of the washboard c 0 10-1 potential disappear at I ≥ I0. If I is decreased from I > I0, the particle becomes retrapped at Ir < I0, lead-

ing to a hysteretic IVC. This behavior can be quantified -2 10 -2 -1 0 1 by the Stewart-McCumber parameter 10 10 β 10 10 2π L β ≡ I R2C. (4) C Φ 0 0 FIG. 2. The dc SQUID: (a) Schematic view. (b) Critical In order to obtain a non-hysteretic IVC, βC must be kept current vs applied magnetic flux for different βL and (c) Ic below ∼ 1. This can be e.g. achievedFig. 2 by means of an modulation vs βL, both calculated for T = 0 and identical additional shunt resistor, parallel to the JJ. JJs. 4

increasing βL [Fig. 2(b,c)]. This effect allows to estimate At higher frequencies, SΦ becomes indepenent of f. L from the measured Ic(Φ). This white noise SΦ,w is mainly due to Johnson-Nyquist We note that the inductance L = Lg + Lk has two noise associated with dissipative quasiparticle currents 28 contributions : The geometric inductance Lg relates in the JJs or shunt resistors. Within a Langevin ap- the induced flux LgJ to the current J circulating in the proach, the thermal noise is described by two indepen- SQUID loop. The kinetic inductance Lk is due to the dent fluctuation terms in the coupled equations of mo- kinetic energy of J and can often be neglected; however, tion for the two RCSJ-type JJs. Numerical simula- it becomes significant when the width and/or thickness tions yield SΦ,w vs βL, βC and the noise parameter 26,30 < of the SQUID ring are comparable to or smaller than the Γ ≡ kBT/EJ = 2πkBT/(I0Φ0) . For βC ∼ 1, βL > 0.4 London penetration depth λL. and ΓβL < 0.1, one finds For most applications, the dc SQUID is operated in Φ k TL the dissipative state as a flux-to-voltage transducer. In S ≈ 4(1 + β ) 0 B . (8) Φ L I R this case, the SQUID is current biased slightly above Ic, 0

leading to a Φ0-periodic modulation of V (Φ), which is < For β ∼ 0.4, S increases again with decreasing β . Typ- often sinusoidal. This mode of operation requires non- L Φ L < ically, SQUIDs are designed to give βL ≈ 1, for which hysteretic IVCs, i.e., β ∼ 1. An applied flux signal δΦ C Eq. (8) reduces to S ≈ 16k TL2/R7. This linear causes then a change δV in SQUID voltage, which for Φ B scaling S ∝ T , however, saturates in the sub-Kelvin small enough signals is given by δV = (∂V/∂Φ) δΦ. Usu- Φ range [Fig. 3(b)] due to the hot-electron effect stemming ally, the working point (with respect to bias current I from limited electron-phonon interaction at low T 36. We and applied bias flux) is chosen such that the slope of √ note that S ∝ L (for fixed β ≈ 1), meaning that the V (Φ) curve is maximum, which is denoted as the Φ L small loop inductances yield lower white flux noise levels. transfer function V = (∂V/∂Φ) . Φ max Other sources of white noise are shot and quantum noise, The sensitivity of the SQUID in the voltage state is lying usually below the Johnson-Nyquist term. For the limited by voltage fluctuations, which are quantified by 7 case βL = 1, the former is given by SΦ ≈ hL , whereas the spectral density of voltage noise power SV . This is the latter arises from zero point quantum fluctuations converted into an equivalent spectral density√ of flux noise 8 2 giving SΦ ≈ hL/π . power√ SΦ = SV /VΦ or the rms flux noise SΦ with units Φ0/ Hz [Fig. 3(a)]. At low frequency f, excess noise scaling typically as C. SQUID readout SΦ ∝ 1/f (1/f noise) shows up. Major sources are critical current fluctuations in the JJs and thermally acti- 1. Flux-locked loop vated hopping of Abrikosov vortices in the superconducting film, which is particularly strong in SQUIDs based on 30 The periodic response of the SQUID to magnetic flux the high-Tc cuprate superconductors . Moreover, 1/f noise has also been ascribed to flux noise arising from can be linearized to obtain a larger dynamic range. This 31 can be achieved by operation in the flux locked loop fluctuating spins at the interfaces of the devices . This 37 is supported by the observation of a paramagnetic sig- (FLL) mode . Here, the SQUID is (typically current) nal following a Curie-like T -dependence32–34. However, biased at an optimum working point and behaves as a a complete description of 1/f noise is still missing. null-detector of magnetic flux. A small variation δΦ of the external flux changes the SQUID output (typically a voltage change δV ). This small deviation from the

C407-R42 working point is ampliﬁed, integrated, and fed back to (a) (b) C401-C43 C509-S33 the SQUID via a current through a feedback resistor Rf ) 10 C509-U33

1/2 and coil, which is inductively coupled to the SQUID. The output voltage across Rf is then proportional to the ﬂux /Hz 0 signal δΦ. The dynamic response in FLL mode is lim- 1 4.2 K µΦ

( ited by the slew rate, i.e. the speed at which the feedback Φ 1/2 circuit can compensate for rapid ﬂux changes at the in- S 13 mK 0.1 put. Under optimum conditions, the bandwidth of the

-1 0 1 2 3 4 FLL is only limited by propagation delays between the 10 10 10 10 10 10 0.01 0.1 1 10 room temperature feedback electronics and the SQUID; f (Hz) T (K) a typical distance of 1 m yields ∼ 20 MHz.

FIG. 3. Rms flux√ noise of Nb thin film SQUIDs with Nb/Al- AlOx/Nb JJs (a) SΦ(f) at 4.2 K and 13 mK [after Mart´ınez- Pérez et al.35] (b) High-frequency (white) noise, measured at 2. Voltage readout different temperatures on different sensors. The white noise Fig. 3 depends on T as expected from theory (SΦ ∝ T ) down to The most simple SQUID readout uses current bi- ∼ 100 mK when it saturates. ased operation in the dissipative state; as mentioned 5 above, the IVCs should be non-hysteretic in this case. on the field direction and temperature41. As the transfer function VΦ is typically small (several 10−100 µV/Φ0), the voltage noise at the output can easily be dominated by room-temperature amplifier noise. 4. Dispersive read out To circumvent this problem, one can use a flux modula- 37 tion scheme . Here, the SQUID is flux-modulated by an So far, we discussed SQUID operation in the volt- ac signal (amplitude Φ0/4, frequency fm ∼ 100 kHz), and age state or close to it. Such schemes entail dissipa- the resulting ac voltage across the SQUID is amplified tion of Joule power that might affect the state of the with a (cold) step-up transformer to increase the SQUID magnetic system under study. An elegant way to cir- signal and noise. The modulated SQUID response is cumvent this problem is the operation of the SQUID further amplified at room temperature and lock-in de- as a flux-dependent resonator; this has also the ad- tected. Suitable electronics achieve a bandwidth of up to vantage of increasing enormously the bandwidth up to 100 kHz. ∼ 100 MHz42,43. The SQUID is always in the supercon- In a different approach, one can increase VΦ by addi- ducting state and acts as a flux-dependent inductance tional positive feedback (APF), which distorts the V (Φ) connected in parallel to a capacitor. The resonance fre- characteristics and increases VΦ at the positive slope. quency of the circuit depends on the total flux threading This enables simple direct readout of the SQUID signal37. the SQUID loop. This can be read out by conventional Alternatively, a low-noise SQUID or serial SQUID array microwave reflectometry giving a direct flux-to-reflected (SSA) amplifier can be used to amplify the SQUID volt- phase conversion. The devices are operated in the linear age at low T in a two-stage readout configuration. regime, i.e., using low-power driving signals. To determine the spectral density of flux noise, the overall voltage noise of the circuit is estimated and scaled with the 3. Critical current readout & threshold detection transduction factor dV/dΦ. The noise performance can be boosted considerably by taking advantage of the CPR For SQUIDs with hysteretic IVCs one can exploit the non-linearity, i.e., operating the nanoSQUID as a parametric amplifier. For this purpose, the driving power is Ic(Φ) modulation directly. In this case one ramps the bias current until the SQUID switches to the dissipative increased so that the resonance peak is distorted, giving state, producing a voltage drop. At this point the current a much sharper dependence of the reflected phase on Φ. is switched off, and Ic is calculated from the duration of the ramp38. This technique can also be used with a FLL scheme38–40. Sensitivity is limited by the accuracy III. NANOSQUIDS: DESIGN, FABRICATION & PERFORMANCE in determining Ic, which is described by the escape of a particle from a potential minimum. Such a process can be thermally activated or quantum driven and is strongly NanoSQUIDs are developed for detecting small spin influenced by electronic noise. Hence, a large number of systems, such as MNPs or SMMs, or for high-resolution switching events is needed to obtain sufficient statistics. imaging of magnetic field structures by SQUID mi- To minimize Joule heating, the SQUID can be oper- croscopy. For such applications, the figure of merit is ated as a threshold sensor. Here, the SQUID is current- the spin sensitivity, which can be boosted down to the biased very close to the switching point. If the magnetic level of a single electron spin. The use of strongly minia- flux threading the loop changes abruptly, the SQUID is turized SQUID loops and JJs is based on the following triggered to the dissipative state and a voltage drop will ideas: be measured38. • Stronlgy localized magnetic field sources Both techniques were applied to magnetization rever- (e.g. MNPs) are placed in close vicinity to the sal measurements on MNPs in sweeping magnetic fields SQUID, instead of using pickup coils [Fig. 4(a)] H38. For measurements up to large H, applied along which degrade the overall coupling. A single any direction, the measurement procedure is divided into SQUID loop [Fig. 4(b)] can be used to detect the three steps. First, H is applied to saturate the particle’s magnetic moment µ of a MNP, or gradiometric magnetization along any direction. Second, H is swept configurations [Fig. 4(c,d)] enable measurements along the opposite direction to a value H and back to test of the magnetic ac susceptibility χ . zero. To check whether this reversed the particle’s mag- ac netization, an in-plane field sweep is done as a third step. • The coupling of the stray field from local field If the particle’s magnetization reversal is (not) detected sources to the nearby SQUID can be improved by in the third step one can conclude that Htest was above reducing the cross section (width and thickness) of (below) the switching field Hsw. These steps can be re- the superconducting thin film forming the SQUID peated several times to determine Hsw precisely. Note loop (see section III A). that the second step can be performed above Tc of the SQUID. Rather than tracing out full M(H) loops, this • The sensitivity of the SQUID to magnetic flux technique can be used to trace out the dependence of Hsw (magnetic flux noise in the thermal white noise 6

(a) SQUID magnetometer (b) nanoSQUID magnetometer Hence, optimizing nanoSQUID performance requires to minimize SΦ while maximizing φµ. B pickup ext µ As mentioned in section II B, SΦ has typically a low- coils frequency 1/f-like contribution and a thermal white noise part SΦ,w. The 1/f contribution is hard to op- µ M timize by design; however, SΦ,w depends on geometrical parameters through the loop inductance L, but also on junction parameters such as I0, R and C. The SΦ(L) B B ac ac dependence (Eq.(8)) implies that SΦ can be improved by µ µ decreasing L via the loop dimensions, while considering the constraints on βC and βL, which will aﬀect the choice of junction parameters. Such an optimization procedure

(c) series-gradiometer (d) parallel-gradiometer can be tested experimentally by performing flux noise measurements of the SQUIDs. FIG. 4. Layouts of various SQUID sensors. (a) SQUID The optimization of the coupling factor φµ = Φ/µ is magnetometer based on gradiometric pickup coils coupled in- more difficult. It is defined as the magnetic flux Φ cou- ductively (via mutual inductance M) to a SQUID. (b)-(d) pled to the SQUID loop by the magnetic dipole field of NanoSQUIDs without intermediate pick-up coils; the stray a point-like particle, divided by its magnetic moment µ. field created by a MNP with magnetic moment µ is directly The magnitude of φµ depends on SQUID geometry, par- sensed by the SQUID loop. Magnetization measurements can ticle position rµ (relative to the SQUID) and orientation be performed by applying an external magnetic field Bext in eˆ = µ/µ of its magnetic moment. This quantity is not the nanoloop plane (b). The frequency-dependent magnetic µ directly accessible by experiments, and one has to rely Fig. 4 ac susceptibility χac can be sensed by using series (c) or parallel (d) planar gradiometers; a homogeneous ac excitation on estimates, analytic approximations or numerical cal- magnetic field Bac is applied perpendicular to the gradiome- culations for determining φµ and optimizing it. 44 ter’s plane through on-chip excitation coils. To the best of our knowledge, Ketchen et al. were the first to give an estimate of φµ. For a magnetic dipole at the center of an infinitely thin loop with radius a, with limit) can be improved by reducing the loop in- eˆµ along the loop normal ductance, i.e. by shrinking the lateral size of the µ SQUID loop (see section III A). φ = 0 = (r /a)·(Φ /µ ) ≈ (2.8 µm/a)·(nΦ /µ ) (9) µ 2a e 0 B 0 B • For magnetization reversal measurements on was found.45 The r.h.s. of Eq. (9) is obtained with the MNPs, an external field Bext is applied ideally ex- 2 µ0e definition of the classical electron radius re = ,Φ0 = actly in the plane of the SQUID loop to switch 4πme h eh µB 2re the MNP’s magnetization (see section IV B), albeit and µB = , which yields = . 2e 4πme Φ0 µ0 without coupling flux directly to the loop. By re- The coupling improves if the particle is moved close ducing the dimensions of the JJs and the loop, the to the loop’s banks46. However, a quantitative estimate 47 nanoSQUID can be made less sensitive to Bext for of φµ is more difficult in this near-field regime , as the small misalignment of Bext. cross-section of the SQUID banks and the flux focusing effect caused by the superconductor must be taken into • Reducing the loop size together with the SQUID- account. The calculation of φ requires calculating the to-sample distance can significantly boost the spa- µ magnetic field distribution at the position of the SQUID, tial resolution for scanning SQUID microscopy ap- originating from a magnetic moment µ at position r , plications (see section V). µ and from this the magnetic flux coupled to the SQUID. This problem can be simplified by exploiting the fact that sources and fields can be interchanged, i.e., one evaluates A. nanoSQUIDs: design considerations the magnetic field BJ (rµ), created by a circulating supercurrent J through the SQUID loop, at the position rµ of the magnetic dipole. With the normalized quantity The ability of a nanoSQUID to resolve tiny signals 46,48 from the magnetic moments of small spin systems de- bJ = BJ /J, which does not depend on J, one finds pends (i) on the intrinsic flux noise S of the SQUID Φ φ (r , eˆ ) = eˆ · b (r ) . (10) and (ii) on the amount of flux Φ which a particle with µ µ µ µ J µ magnetic moment µ couples to the SQUID loop. The lat- This allows to calculate φµ for any position and orien- ter can be quantified by the coupling factor φ ≡ Φ/µ, µ tation of the magnetic dipole in 3D space once bJ is with µ ≡ |µ|. As a result, one can define the spin known.49 p √ √ sensitivity Sµ = SΦ/φµ, with units µB/ Hz; µB is The normalized field bJ has to be determined from p the Bohr magneton. Sµ expresses the minimum mag- the spatial distribution of the supercurrent density js netic moment that can be resolved per unit bandwidth. circulating in the SQUID loop, which depends only on 7

coupled to a ferromagnetic Fe tip, which was scanned over the sensor’s surface while recording the SQUID output in open-loop configuration57. The optimization of the spin sensitivity in the thermal white noise limit requires the knowledge of the dependence of φµ and SΦ,w on SQUID geometry, as this affects both the SQUID inductance and the coupling. A detailed investigation of this problem was done for YBCO nanoSQUIDs55 (see section III C). This study shows that it is essential to consider the increase in kinetic inductance Lk when the thickness and width of the loop is reduced to a length scale comparable to or even smaller than λL. Hence, to improve the Sµ one has to find a com- promise between improved coupling and deterioration of flux noise (via an increased Lk) upon shrinking the cross section of the SQUID loop.

B. nanoSQUIDs based on metallic superconductors

1. Sandwich-type SIS junctions FIG. 5. Calculated coupling factor φµ vs position of a magnetic dipole pointing in x-direction on top of Nb nanoSQUIDs. Main graphs show contour plots φµ(x, z) for (a) a magne- The SIS junction technology (S: superconductor, I: tometer and (b) a gradiometer. Nb structures are indicated thin insulating barrier), typically producing JJs in a by black rectangles; dashed lines indicate position of lines- Nb/Al-AlOx/Nb trilayer geometry, is the most com- cans φµ(x) [above (a)] and φµ(z) [right graphs]. Insets show monly used approach to fabricate conventional SQUID- scanning electron microscopy (SEM) images. Reprinted with based devices. This technology is highly developed and Fig. 5 50 permission from Nagel et al. . Copyright [2011], AIP Pub- reproducible, yielding high-quality JJs with controllable lishing LLC. critical current densities jc from ∼ 0.1 up to a few kA/cm2 at 4.2 K. However, a major disadvantage is the low jc, which results in too small values for the critical the SQUID geometry and on λL. This has been done current if submicron JJs are used. As a consequence, for various types of nanoSQUIDs by numerically solv- even if the SQUID loops are miniaturized, the operation ing the London equations48,50–55. Numerical simula- of micron-sized JJs in large magnetic fields is only pos- tions of φµ reveal that the coupling can be increased in sible with careful alignment of the field perpendicular to the near-field regime if the magnetic dipole is placed as the junction plane, as an in-plane field in the 1-10 mT close as possible on top of a constriction in the SQUID range can easily suppress the critical current due to the 55 loop, which is as thin and narrow as possible . Typi- Fraunhofer-like modulation of Ic(B). Frequently used cal φµ = 10 − 20 nΦ0/µB have been obtained for mag- window-type JJs come with a large parasitic capacitance netic dipoles at 10 nm distance from a constriction (∼ due to the large area of surrounding superconducting lay- 100 − 200 nm wide and thick) in YBa2Cu3O7 (YBCO) ers. A commonly used approach is therefore to use nor- nanoSQUIDs.56 Simulation results for two types of Nb mal metal layers to shunt these junctions, for lowering nanoSQUIDs [Fig. 5] show that the dipole has to ap- βC to yield non-hysteretic IVCs, albeit at the cost of proach the SQUID surface closely to reach values above also lowering the characteristic voltage Vc = I0R. The a few nΦ0/µB (see φµ(z) linescans in the right graphs absence of hysteresis offers the advantage to operate the in Fig. 5). The φµ(x) linescans (top graph in Fig. 5) SQUID as a flux-to-voltage converter, using conventional show that the coupling is maximum right above the loop readout techniques. structures50. As a key advantage, the SIS technology offers a well Measurements on spatially extended magnetic sys- developed multilayer process, allowing for the realiza- tems, such as a Ni nanotube51 or a Fe nanowire53, were tion of more complex designs, as compared to a sin- found to be consistent with the numerical approach de- gle layer technology. This allows for the fabrication of scribed above. This was done by comparing the measured superconducting on-chip input circuits such as coupling flux coupled to nanoSQUIDs from fully saturated tubes transformers, susceptometers or advanced gradiometers. or wires with the calculated flux signals, obtained by in- This approach has been taken very successfully to real- tegrating φµ over the finite volume of the sample. First ize miniaturized structures for applications in magnetic measurements on the SQUID response as a function of particle measurements and scanning SQUID microscopy, the position of a magnetic sample have been reported although those did not really involve SQUIDs with (lat- earlier. In those experiments, small SQUID sensors were eral outer) dimensions in the submicrometer range. 8

field coils

FIG. 6. SEM image of a SQUID microsusceptometer which a nanoloop patterned in the pickup coil (inset). Images courtesy of J. 100 µm Ses´e.

The first SQUID device designed to measure mag- Submicrometric Nb/AlOx/Nb JJs in a cross-type de- netic signals from MNPs was based on micrometric sign were recently used for fabricating miniaturized 64 Nb/NbOx/Pb edge junctions, which were connected in SQUIDs . The key advantage of cross-type JJs over parallel to two oppositely wound loops to form a micro- conventional window-type JJs is the elimination of the 9 susceptometer√ . The white flux noise at 4.2 K was parasitic capacitance surrounding the JJ, which becomes 0.84 µΦ0/ Hz. This susceptometer was operated in a increasingly important upon reducing the JJ size. At dilution refrigerator, and the output signal was mea- T = 4.2 K, 0.8 × 0.8 µm2 JJs show non-hysteretic IVCs, sured in open-loop configuration and amplified by an rf if they are shunted with a AuPd layer. Sensors are SQUID preamplifier. Magnetic susceptibility measure- also produced with an integrated Nb modulation coil. ments performed with this system will be reviewed in sec- Square-shaped washer SQUIDs with minimum inner size tion IV C. Very similar devices based on Nb/Al-AlO /Nb of 0.5 µm have an inductance of a few pH. SQUIDs oper- √ √ x√ JJs with SΦ = 0.8 µΦ0/ Hz at 4 K and 0.25 µΦ0/ Hz ated in liquid He and read out with√ a low-noise SQUID p 65 Fig. 6 below 0.5 K were adapted to the use in scanning SQUID preamplifier yield SΦ,w = 66 nΦ0/ Hz . microscopes58,59; see section V. Broad-band SQUID microsusceptometers have been realized by locally modifying SQUID current sen- 2. Sandwich-type SNS junctions sors based on Nb/Al-AlOx/Nb JJ technology. Those sensors60 come in two types: (i) high-input inductance SNS junctions (N: normal conductor) offer the advan- > 5 2 (∼ 1 µH) sensors incorporate an intermediate trans- tage of large critical current densities ∼ 10 A/cm at former loop with gradiometric design; (ii) low-input in- 4.2 K and non-hysteretic IVCs, albeit at the cost of some- ductance (2 nH) devices without intermediate loop; here what reduced I0R values. Hence, this type of JJs is very the input signal is directly coupled to the SQUID via well suited for fabricating nanoSQUIDs with junction size four single-turn gradiometric coils connected in parallel. in the deep submicron range. These SQUIDs are non-hysteretic down√ to sub-K tem- In a Nb/HfTi/Nb trilayer process, originally devel- p 66 peratures with SΦ,w = 800 nΦ0/ Hz at T = 4.2 K. oped for Josephson arbitrary waveform synthesizers , Modification of these sensors was done by FIB milling JJs with 200 × 200 nm2 area or even below are obtained and FIB-induced deposition (FIBID) of superconduct- by e-beam lithography and chemical-mechanical polish- 35,61 50,54 ing material with W(CO)6 as precursor gas . This ing, producing nanoSQUIDs with 24 nm thick HfTi allowed converting the intermediate transformer loop barriers; the latter can be varied to modify jc. As for the into a susceptometer inductively coupled to the SQUID SIS JJ technology, the fabrication process offers much [Fig. 4(a)]. By modifying the gradiometric microSQUID flexibility for realizing complex designs. Both series- itself it is possible to directly couple an MNP to the and parallel-gradiometers and single SQUID loops were SQUID loop34 [Fig. 4(d)]. Later, SQUID-based micro- realized50,54,67. Devices were patterned in a washer- or susceptometers with improved reflection symmetry were microstrip-type geometry, with the loop plane parallel produced62,63. The sensitivity was boosted by defining a or perpendicular to the junction’s (substrate) plane, re- nanoloop (450 nm inner diameter, 250 nm linewidth) by spectively. A key advantage of the microstrip-type ge- FIB milling in one of the pickup coils [Fig. 6]. These sen- ometry [Fig. 7] is the possibility to realize very small sors offer an extremely wide bandwidth (1 mHz - 1 MHz) loop areas, defined by the thickness of the insulating in- and can be operated at T = 0.013 − 5 K for the inves- terlayer between the top an bottom Nb lines times the tigation of microscopic crystals of SMMs and magnetic lateral separation of the two JJs. This results in very proteins; such measurements will be reviewed in section small SQUID inductances, typically a few pH. Moreover, IV C. a magnetic field applied in the plane of the SQUID loop 9

HfTi Nb 2 600 nm loop size: 600 × 220 nm HfTi Nb

z Imod Φmod y x

FIG. 7. Layout of Nb/HfTi/Nb nanoSQUID in microstrip geometry. Arrows indicate flow of bias current I, modulation current Imod and direction of external field B. Inset shows SEM image with JJs (200 × 200 nm2) indicated by dashed Fig. 9 FIG. 9. SEM image of a 3-dimensional nanoSQUID fabricated squares. SEM image courtesy of B. Müller. Fig. 7 using FIB sculpting and all Nb technology. The flux capture area of the nanosensor is 1 × 0.2 µm2, and the two Josephson tunnel junctions have an area of about 0.3×0.3 µm2. The inset can be perpendicular to the JJ (and substrate) plane; is a sketch of the device, showing the current paths through 69 in this way the field-induced suppression of Ic can be the device. Reprinted with permission from Granata et al. . avoided. It has been shown that magnetic fields up to Copyright [2013], AIP Publishing LLC. 0.5 T can be applied while degrading only marginally the 54 performance . On-chip flux biasing is easily possible√ for operation in FLL. White flux noise ∼ 110 nΦ0/ Hz has shown in Fig. 8. Two microstrip-type Nb nanoSQUIDs been obtained. Based on numerical solutions of the Lon- SQx and SQy, as described above, with perpendicular don equations for φµ, this yields a spin sensitivity of just loops are sensitive to fields in x- and y-direction, respec- √ tively. A third SQUID, SQz has a gradiometric layout, ∼ 10 µB/ Hz for a magnetic dipole 10 nm away from the SQUID loop. Magnetization measurements on magnetic in order to strongly reduce its sensitivity to the applied nanotubes have been performed successfully and will be homogeneous magnetic field. Simultaneous operation of summarized in section IV B. all three nanoSQUIDs in such devices in FLL has been demonstrated at 4.2 K in fields up to 50 mT, with a flux By combining three mutually orthogonal nanoSQUID √ 1/2 < loops, a 3-axis vector magnetometer has been realized noise SΦ,w∼ 250 nΦ0/ Hz. By numerical simulations of very recently68. Here, the idea is to distinguish the the coupling factor, it has been demonstrated that for three components of the vector magnetic moment µ of a MNP placed in the center of the left loop of the gra- a MNP placed at a specific position, and subjected to an diometer (cf. Fig. 8), the three orthogonal components of applied magnetic field along z-direction for magnetiza- the magnetic moment of the MNP can be detected with a tion reversal measurements. The layout of the device is relative error flux below 10 %. Such a device can provide important information on the magnetic anisotropy of a single MNP. Submicrometer nanoSQUIDs have recently also been 500 nm fabricated based on SNIS JJs69. Starting from a Nb/Al- AlOx/Nb trilayer, a three dimensional SQUID loop 2 SQx (0.2 µm ) was nanopatterned by FIB milling and an- SQy odization [Fig. 9]. The resulting JJs have an area of approximately 0.3 × 0.3 µm2 and are intrinsically shunted by the relatively thick (80 nm) Al layer, yielding non- hysteretic IVCs. The smallness of the SQUID loop leads to L = 7 pH. Measurements at 4.2 K yield pS ∼ √ Φ,w SQz 0.68 µΦ0/ Hz.

3. Constriction junctions FIG. 8. SEM image of a 3-axis vector magnetometer, consisting of two orthogonal nanoSQUIDs (SQx, SQy) and an Josephson coupling can also occur in superconducting z 70 Fig.orthogonal 8 gradiometric nanoSQUID (SQ ). Black dotted constrictions (Dayem bridges ) with size similar to or squares indicate positions of Josephson junctions. smaller than the coherence length ξ(T )23. The IVCs 10

FIG. 10. cJJ-based nanoSQUIDs: (a) schematic view with a MNP (magnetic (a) (b) moment µ) close to one con- µ striction where coupling is Bext maximum. (b) SEM images of Nb microSQUID with Ni wire on top (left) and Nb nanoSQUID (right), drawn to scale in left graph. Graph (b) Reproduced with permission from38. All rights reserved c IOP Publishing [2009]. of such constriction-type Josephson junctions (cJJs) are Al and 1 T for Nb. From the measured critical√ current often hysteretic, due to the heat dissipated above Ic. noise, the flux noise√ was calculated as ∼ 40 µΦ0/ Hz for 71 Short enough cJJs show a sinusoidal CPR; however, a Al and ∼ 100 µΦ0/ Hz for Nb . Due to hysteretic IVCs significant deviation occurs if the constriction length is these nanoSQUIDs were operated in Ic readout mode or larger than ξ, which can even lead to multivalued CPRs. as threshold detectors (see section II C 3). These sensors Hence, optimization of SQUID performance based on an allowed the vastest realization of true magnetization mea- RCSJ analysis is difficult, and hysteretic IVCs prevents surements (section IV C) and were also implemented into conventional SQUID operation with current bias. Still, probe tips to perform scanning SQUID microscopy71,73. non-hysteretic IVCs can be achieved by operation close For similar Nb cJJ-based nanoSQUIDs (30 nm thick, enough to Tc, were Ic is reduced, or by adding a metal- ∼ 200 nm inner loop size, cJJs down to 280 nm long and lic overlayer as a resistive shunt. Another drawback is 120 nm wide) switching current distributions were mea- theFig. large 8 kinetic inductance Lkin of the constriction, that sured from 4.2 down to 2.8 K74. A detailed analysis of the can dominate the total SQUID inductance L and prevent noise performance for Ic readout revealed a flux sensitiv- improving the flux noise by shrinking the loop size. On ity of a few mΦ , which was shown to arise from ther- the other hand, cJJ-based nanoSQUIDs in a simple pla- 0 mally induced Ic fluctuations in the nanobridges. More nar configuration can be fabricated relatively easily from recently, hysteretic nanoSQUIDs made of Al-Nb-W lay- thin film superconductors, e.g., Al, Nb or Pb, through ers (2.5 µm inner loop size; 40 nm wide, 180 nm long cJJs) one-step electron-beam (e-beam) or FIB nanopattern- could be operated with oscillating current-bias and lock- ing. Moreover, the use of nanometric-thick films and the 75 in read-out at T < 1.5 K . In this configuration Ic is smallness of the constriction makes these SQUIDs quite considerably reduced due to the inverse proximity effect insensitive to in-plane magnetic fields and yields large of W on Nb. coupling factors if MNPs are placed close to the constriction [Fig.10(a)]. The small size of cJJs is a key advantage Nanometric Nb SQUIDs (50 nm thick, down to 150 nm for fabricating nanoSQUIDs with high spin sensitivity. inner hole size) were also fabricated by FIB milling to produce cJJs (80 nm wide, 150 nm long)76. It was ob- First thin film Nb dc SQUIDs based on cJJs with served that Ga implantation depth can reach values of linewidths down to 30 nm, patterned by e-beam lithog- 30 nm, suppressing the superconducting properties of Nb. raphy, were reported in 19805. Despite their large L = At T = 4.2 K, devices with relatively small I < 25 µA 150 pH, miniaturized SQUIDs, with loop size ∼ 1 µm2, c √ showed nonhysteretic IVCs and could be operated in exhibited low flux noise ∼ 370 nΦ / Hz at 4.2 K. Dur- 0 a conventional current-bias mode, yielding pS ∼ ing the 1990s, the use of cJJ nanoSQUIDs for the in- √ Φ,w vestigation of small magnetic systems was pioneered by 1.5 µΦ0/ Hz. Wernsdorfer et al.11,12,38. Figure 10(b) shows examples A possible way to approach the sinusoidal CPR of of such devices, which were patterned by e-beam lithog- ideal point contacts is the use of variable thickness raphy from Nb and Al films71. Typical geometric pa- nanobridges. Here, the thicker superconducting banks rameters were 1 µm2 inner loop area, 200 nm minimum can serve as phase reservoirs, while the variation in the linewidth and 30 nm film thickness. The size of the con- superconducting order parameter should be confined to strictions (∼ 30 nm wide, ∼ 300 nm long) was signifi- the thin part of the bridges77. cJJ-based nanoSQUIDs cantly larger than ξ for Nb. This lead to a highly non- were realized by local anodization of ultrathin (3−6.5 nm- 23,72 ideal CPR and hence non-ideal Ic(Φ) dependence thick) Nb films using a voltage-biased atomic force micro- 78 with strongly suppressed Ic modulation depth for Nb cJJ scope (AFM) tip . This technique produced constric- SQUIDs. Furthermore, Lkin of the constrictions can be tions (30 − 100 nm wide and 200 − 1000 nm long) and a few 100 pH, dominating the overall inductance of the variable thickness nanobridges by further reducing the devices72. Impressively large magnetic fields could be constriction thickness down to few nm (within a ∼ 15 nm applied parallel to the nanoSQUID loops up to 0.5 T for long section). The latter exhibited ∆Ic/Ic twice as large 11 as the former, indicating an improved CPR. of the loop, which is limited by the upper critical field of Vijay et al.79 produced Al nanoSQUIDs based on the superconductors. The use of very thin superconduct- cJJs (8 nm thick, 30 nm wide) with variable length (l = ing layers can increase the effective critical field. Follow- 75 − 400 nm). The cJJs were either connected to super- ing this idea, 3 − 5 nm-thick cJJ Nb nanoSQUIDs were conducting banks of the same thickness (‘’2D devices”) fabricated, supporting in-plane fields up to 10 T. These or to much thicker (80 nm) banks (‘’3D devices”). For 3D sensors proved to be well suited for measuring magnetiza- 89 devices with l ≤ 150 nm ≈ 4ξ, the measured Ic(Φ) curves tion curves of microcrystals of Mn12 SMMs . However, indicate a CPR which is close to the one for an ideal their large kinetic√ inductances lead to large flux noise short metallic weak link. Both 2D and 3D devices were (∼ 100 µΦ0/ Hz). More promising is the use of materi- fully operative up to in-plane magnetic fields of 60 mT80. als with larger upper critical fields, such as boron-doped Such nanoSQUIDs were operated with dispersive readout diamond90. Micrometric SQUIDs based on 100 nm-wide constrictions in 300 nm thick films were demonstrated to (see section√ II C 4) yielding impressive flux noise values 43 of 30 nΦ0/ Hz for a 20 MHz bandwidth . operate up to impressive fields of 4 T applied along any Variable thickness bridges have recently also been re- direction. These devices were, however, hysteretic due to alized by connecting suspended Al nanobridges (25 nm heat dissipation. Flux sensitivity was determined√ from thick, 233 nm long, 60 nm wide) to Nb(30 nm)/Al(25 nm) the critical current uncertainty giving 40 µΦ0/ Hz. bilayer banks to form a nanoSQUID (2.5 µm-in-diameter Finally we note that the smallest nanoSQUIDs real- loop)81. These devices have the advantage of using cJJs ized so far, which also include cJJs, are the SQUIDs-on- 91,92 from a material (Al) with relatively large ξ, while main- tip (SOTs) . These devices will be discussed in more taining relatively high Tc and critical magnetic field in detail in section V. the superconducting banks forming the SQUID loop. Thermal hysteresis in the IVCs of cJJs can be suppressed by covering the devices with a normal metal- 4. Proximized structures lic layer, which provides resistive shunting and acts as a heat sink. cJJ-based nanoSQUIDs from 20 nm-thick A normal metal in good contact between supercon- Nb films covered by 25 nm-thick Au have been patterned ducting electrodes acquires some of their properties due by e-beam lithography to realize 200 nm inner loop size to the proximity effect, inducing a mini-gap in the density and constriction widths in the range 70 − 200 nm, yield- of states of the normal metal. Andreev pairs can prop- ing L ∼ 15 pH82 . The Au layer prevented hysteresis agate along relatively long distances at low T , carrying in the IVCs at temperatures above 1 K, allowing con- information on the macroscopic phase of the supercon- ventional SQUID readout in the voltage state, yielding ductor. In the long (short)-junction regime, when the p √ SΦ,w ∼ 5 µΦ0/ Hz at 4.2 K, increasing by about 15 % Thouless energy of the metal is larger (smaller) than the when operating in a magnetic field of 2 mT83. Field op- superconducting energy gap, the junction properties will eration up to few 100 mT was improved by reducing the be governed by the normal metal (superconductor). hole size down to 100 nm and the largest linewidths down The first dc SQUID built with long proximized JJs to 250 nm84. Preliminary experiments were performed on was based on a CNT intersecting an Al ring93.A ferritin nanoparticles attached to the cJJs85. However, gate-modulated supercurrent was demonstrated and flux- the magnitude of the flux change observed in some cases induced modulation of the critical current (few nA) was (up to 440 µΦ0) was larger than the expected one for a observed at mK temperatures. The goal was to exploit 2 ferritin NP located at optimum position (up to 100 µΦ0). the small cross section of the CNT (∼ 1 nm ) to pro- Low-noise nanoSQUIDs from a Nb/amorphous W bi- vide optimum coupling for molecular nanomagnets at- layer (200 and 150 nm thick, respectively) have been pro- tached to it. An experimental proof-of-principle of such duced by FIB milling86. The SQUID loop (370 nm in- a CNT-based magnetometer is, however, still missing. A ner diameter) was intersected by two nanobridges (65 nm micrometric dc SQUID with graphene proximized junc- wide and 60 − 80 nm long) which showed non-hysteretic tions (50 nm long, 4 µm wide) was also reported94. Flux- IVCs at 5 − 9 K. Readout in the voltage state gave induced Ic modulation was observed, however, no noise p √ SΦ,w = 200 nΦ0/ Hz at 6.8 K. Recently, the same performance of the device was reported. group extended the operation temperatures down to < Micrometric dc SQUIDs containing normal metal 1 K by using superconducting Ti films, inversely prox- bridges as weak links have also been reported. 87 imized by Au layers to reduce Tc . These SQUIDs Nb/Au/Nb and Al/Au/Al-based devices showed IVCs (with 40 nm wide and 120 nm long constrictions) exhib- with pronounced hysteresis, due to heat dissipated in the ited no hysteresis within 60 mK < T < 600 mK and had normal metal after switching95. SQUIDs with shorter p √ SΦ,w = 1.1 µΦ0/ Hz. These devices allowed the de- Cu nanowires (280 − 370 nm long, 60 − 150 nm wide, tection of the magnetic signal produced by a 150 nm di- 20 nm thick) enclosed in a V ring were non-hysteretic. 7 ameter FePt nanobead having 10 µB at 8 K in fields up NanoSQUIDs based on proximized InAs nanowires (∼ to 10 mT88. 90 nm diameter, 20 or 50 nm long) were also reported96 As mentioned earlier, cJJ-based nanoSQUIDs can be with JJs in the intermediate length regime [Fig.11]. operated in strong magnetic fields applied in the plane A different kind of interferometer consists of a su- 12

SQUIPTs are in their early stage of development100, still showing a very narrow temperature range of operation limited to sub-kelvin. On the other hand, they exhibit record low dissipation power of just ∼ 100 fW (Ic ∼pA, Vout ∼ 100 mV)√ and should achieve ﬂux noise levels of just a few nΦ0/ Hz. The latter has not been determined experimentally yet due to limitations from the voltage noise of the room temperature ampliﬁers.

µ 1 m C. NanoSQUIDs based on cuprate superconductors

FIG. 11. SEM image of a SQUID sensor consisting of a prox- High-Tc cuprate superconductors such as YBCO have imized highly doped InAs nanowire enclosed within a V ring very small and anisotropic values of ξ, reaching ∼ 1 nm [after Spathis et al.96]. SEM image courtesy of F. Giazotto for the a − b plane and a minute ∼ 0.1 nm for the c axis, and S. D’Ambrosio. making the fabrication of cJJs extremely challenging. Still, the fabrication of YBCO cJJs with 50 nm × 50 nm Fig. 11 cross-section and 100 − 200 nm length has been reported 101 perconducting loop interrupted by a normal metal is- recently . These JJs exhibit large Ic of a few mA at land. A magnetic field applied to the loop varies the 300 mK. NanoSQUIDs based on this technology were fab- phase difference across the normal metal wire, allowing ricated and preliminary measurements showed low flux p √ flux-modulation of the minigap. This behavior can be noise SΦ,w = 700 nΦ0/ Hz at 8 K. probed by an electrode tunnel-coupled to the normal Probably the most mature JJs from cuprate supercon- metal island [Fig.12], providing a flux-modulated elec- ductors are based on Josephson coupling across grain tric response similar to conventional dc SQUIDs. This boundaries (GBs). Grain boundary junctions (GBJs) device received the name Superconducting Quantum In- can be fabricated, e.g., by epitaxial growth of cuprate su- terference Proximity Transistor (SQUIPT), for being the perconductors on bicrystal substrates or biepitaxial seed magnetic analog to the semiconductor field-effect transis- layers102–104. Although micrometric SQUIDs based on tor. SQUIPTs were pioneered by Giazotto et al.97 using GBJs have been produced30, the miniaturization of high- Al loops and Cu wires (∼ 1.5 µm long, ∼ 240 nm wide). quality GBJs is challenging, because of degradation of These magnetometers were further improved by reducing the material due to oxygen loss during nanopatterning. the length of the normal metal island down to the short- NanoSQUIDs made of high-Tc GBJs are, on the other junction limit, leading to a much larger mini-gap opening. hand, very attractive due to their large critical current By choosing proper dimensions of the normal metal is- densities (∼ 105 A/cm2 at 4.2 K) and huge upper critical land, such sensors do not exhibit any hysteresis down to fields (several tens of T). mK temperatures98,99 and can be voltage or current bi- YBCO GBJ nanoSQUIDs were fabricated by FIB 48,52,53 ased, providing impressive values of VΦ of a few mV/Φ0. milling . Devices consist of 50 − 300 nm thick YBCO epitaxially grown on bicrystal SrTiO3 substrates (24◦ misorientation angle) and covered by typically 60 nm

GB clean barriers S

tunnel N barrier

200 nm FIG. 13. SEM image of YBCO nanoSQUID loop (400 × FIG. 12. Scheme of a SQUIPT. The inset shows a SEM im- 300 nm2), intersected by 130 nm wide GBJs; the GB is indi- age of the SQUIPT core; a normal metal probe is tunnel- cated by the vertical dashed line. The loop contains a 90 nm connected to a proximized Cu island enclosed within an Al wide constriction for ﬂux biasing and optimum coupling. [af- ring. SEM image courtesy of F. Giazotto and S. D’Ambrosio. ter Schwarz et al.52]

Fig. 12

Fig. 13 13

thick Au serving as resistive shunt and to protect the the extremely low white noise level, 1/f-like excess noise YBCO during FIB milling. Typical inner hole size is dominates the noise spectrum within the entire band- 200 − 500 nm and GBJs are 100 − 300 nm wide [Fig. 13]. width of the readout electronics. Bias reversal can only Devices are non-hysteretic and work from < 1 K up to partially eliminate this excess noise, which deserves fur- ∼ 80 K. Large magnetic fields can be applied perpendic- ther investigation. ular to the GBJs in the substrate plane, without severe Finally, an encouraging step towards the controlled for- 52 degradation of the Ic modulation for fields up to 3 T . mation and further miniaturization of high-Tc JJs has 105 Via a modulation current Imod through a constriction been made recently by . For this purpose a 0.5-nm- (down to ∼ 50 nm wide) in the loop, the devices can be diameter He+-beam was used to fabricate ∼ 1 nm-narrow flux-biased at their optimum working point, without ex- ion-irradiated barriers on 4 µm wide and 30 nm thick ceeding the critical current, i.e. the constriction is not YBCO bridges. The key point is the smallness of the ion acting as a weak link. The constriction is also the posi- beam diameter, which allows the introduction of point tion of optimum coupling of a MNP to the SQUID. like defects. By varying the irradiation dose between + Numerical simulations based on London equations for 1014 − 1018 He /cm2 the authors showed the successful variable SQUID geometry provided expressions for L and realization of JJs exhibiting SNS-like or tunnel-like be- φµ [via Eq. (10)] for a magnetic dipole 10 nm above the havior. This technique has been applied to the fabrica- constriction, as a function of all relevant geometric pa- tion of SQUID devices106, but their downsizing to the rameters. Together with RCSJ model predictions for nanoscale still needs to be realized. SΦ,w at 4.2 K, an optimization study for the spin sensitivity has been performed. An optimum film thickness dopt = 120 nm was found (for λL = 250 nm). For IV. NANOSQUIDS FOR MAGNETIC PARTICLE smaller d, the increasing contribution of Lkin to the flux DETECTION noise dominates over the improvement in coupling. For optimum βL ∼ 0.5 and d = dopt, the spin sensitivity de- Originally, nanoSQUIDs were conceived for the inves- creases monotonically with decreasing constriction length tigation of individual MNPs and SMMs. These sys- lc (which fixes the optimum constriction width wc). For tems are of key technological importance with appli- lc and wc of several√ tens of nm, an optimum spin sensi- cations ranging from electronics, including hard discs, tivity of a few µB/ Hz was predicted in the white noise magnetic random access memories, giant magneto re- limit55. sistance devices, and spin valves, through on-chip adia- For an optimized device with small inductance L ∼ batic magnetic coolers, and up to biotechnology appli- 4 pH (d = 120 nm, lc = 190 nm, wc = 85 nm), direct cations including enhanced imaging of tissues and or- readout measurements of the magnetic flux noise at 4.2 K gans, virus-detecting magnetic resonance imaging, and √ 107 gave 50 nΦ0/ Hz at 7 MHz (close to the intrinsic thermal cancer therapy (see, e.g., Ref. ). Moreover, magnetic noise floor), which is amongst the lowest values reported molecules appear as an attractive playground to study for dc SQUIDs so far [Fig. 14]. With a calculated cou- quantum phenomena108 and could eventually find application in emerging fields of quantum science such as solid- pling factor φµ = 13√ nΦ0/µB, this device yields a spin 53 state quantum information technologies109 and molecular sensitivity of 3.7 µB/ Hz at 7 MHz and 4.2 K . Due to spintronics110. In this section we will review, as an important application of nanoSQUIDs, the investigation of small magnetic particles. We will first address challenges and approaches experiment regarding positioning of MNPs close to the SQUIDs and ) 10 fit

1/2 then discuss measurements of magnetization reversal and of ac susceptibility of MNPs. /Hz

0 T=4.2 K Φ 1 (µ A. Nanoparticle positioning 1/2 Φ S 0.1 The manipulation and positioning of MNPs close to Φ 1/2 the nanoSQUIDs is particularly important since the 45 n 0/Hz 0.02 magnetic signal coupled to any form of magnetometer 100 101 102 103 104 105 106 107 strongly depends on the particle location with respect to f (Hz) the sensor. Although conceptually very simple, this problem has hampered the realization of true single-particle FIG. 14. (b) Rms flux noise of optimized YBCO nanoSQUID, magnetic measurements so far. Many strategies have measured in open-loop mode. Dashed line is a fit to the mea- been developed to improve the control on the position- sured spectrum; horizontal line indicates fitted white noise. 53 ing of MNPs or SMMs on specific areas of nanoSQUID [after Schwarz et al. ] sensors.

Fig. 14 14

(a) (b)

FIG. 15. SEM images of (a) Co nanoparticle deposited by FEBID on the constriction of a YBCO nanoSQUID and (b) nanodot deposited by FIBID on a SiNi cantilever above a Nb nanoloop. Par- 200 nm 1 µm ticles are highlighted by dashed circles SEM images courtesy of J. Ses´e.

1. In-situ nanoparticle growth SEMs have also been used for this purpose. For instance, a sharpened carbon fiber mounted on a micromanipulator In an early approach, called the drop-casting method, in a SEM has been used to pick up a ∼ 0.15 µm diameter 88 small droplets with suspended MNPs were deposited on single FePt particle an deposit it onto a nanoSQUID . a substrate containing many nanoSQUIDs. After sol- Alternatively, larger carriers that are more easily visi- vent evaporation some of the MNPs happened to oc- ble and manipulated can be used to manipulate the posi- cupy positions of maximum coupling. This method was tion of MNPs. For example, microscopic SiNi cantilevers successfully applied to investigate 15 − 30 nm individual containing the MNP of interest can be moved using a Fig. 13 118 Co MNPs111 . In a similar approach, MNPs based on micromanipulator [Fig. 15(b)]. In particular, CNTs Co, Fe or Ni were sputtered using low-energy cluster appear as promising tools for this purpose. SMMs have beam deposition techniques onto substrates containing indeed been successfully grafted over or encapsulated in- a large amount of microSQUIDs112. Alternatively, MNP side CNTs, which were later used to infer their magnetic and Nb deposition was realized simultaneously to em- properties17. Similarly, an Fe nanowire encapsulated in bed nanometric clusters into the superconducting films, a CNT has been mounted by micromanipulators on top which were subsequently patterned to form nano- or of YBCO nanoSQUIDs for magnetization reversal mea- microSQUIDs113. The drawback of these techniques is surements (see section IV B)53. the lack of precise control of the MNP positions relative Another promising approach is dip pen nanolithogra- to the SQUIDs, which requires the use and characteriza- phy (DPN). Here, an AFM tip is first coated with a solution of many tens or even hundreds of SQUIDs. tion containg MNPs and then brought into contact with Improved nanometric control over the particle po- a surface at the desired location. Capillarity transport sition can be achieved by nanolithography methods. of the MNPs from the tip to the surface via a water This has been used to define Co, Ni, TbFe3 and meniscus enables the successful deposition of small collec- 119 Co81Zr9Mo8Ni2 MNPs with smallest dimension of 100 × tions of molecules in submicrometer dimensions . Bel- 50 × 8 nm311. Alternatively, focused e-beam induced lido et al.120 showed that this technique can be applied deposition (FEBID) of high-purity cobalt (from a pre- to the deposition of dot-like features containing mono- 114 cursor gas, e.g., Co2(CO)8 ) allows the definition of layer arrangements of ferritin-based molecules onto mi- much smaller particles (down to ∼ 10 nm) and arbitrary croSQUID sensors [Fig. 16(a)] for magnetic susceptibility shape located at precise positions with nanometric reso- measurements121 (section IV C). The number of MNPs lution. This technique has been successfully applied to deposited per dot can be controlled (via the concentra- the integration of amorphous Co nanodots onto YBCO tion of the ferritin solution and dot size) from several hun- nanoSQUIDs [Fig. 15(a)]115. dred of proteins down to individual ones120. Recently, DPN has also been applied to the deposition of dot-like features containing just 3−5 molecular layers of Mn12 and 2. Scanning probe-based techniques Dy2 SMMs onto the active areas of microSQUID-based susceptometers, enabling the detection of their magnetic 122,123 The use of a scanning probe, e.g., the tip of an AFM susceptibility [Fig. 16(b)]. can be used for precise manipulation of the position of a Recently, individual magnetic nanotubes, attached to MNP. AFM imaging in non-contact mode is first used to an ultrasoft cantilever were brought in close vicinity to locate MNPs dispersed over a surface Then, using con- a nanoSQUID at low T 51,124,125. This technique allowed tact mode, the tip is used to literally ‘’push” the MNP to the authors to investigate magnetization reversal of the the desired position116,117. This technique was applied to nanotubes by combining torque and SQUID magnetom- improve the coupling between a nanoSQUID and Fe3O4 etry (see section IV B). NPs (15 nm diameter) deposited via the drop-casting We note that scanning SQUID microscopy could also method38. Micro- and nanomanipulators installed inside be applied to the study of MNPs deposited randomly on 15

(a) (b) 5 µm

7 µm

FIG. 16. (a) Ferritin nanodots (dashed circles) deposited by DPN on top of the pickup coil of a SQUID-based microsusceptometer. Each dot contains 104 proteins approximately arranged as a monolayer. Scheme of the DPN nanopatterning technique; a conventional AFM probe delivers dot-like features containing monolayer arrangements of ferritin over the surface [after Mart´ınez-Pérez et al.121]. (b) Optical microscope image taken during the DPN patterning process showing the AFM probe over a microsusceptometer’s pickup coil. The blow-up shows an AFM image of the resulting sample containing 5 molecular layers of Dy2 SMMs. Images courtesy of F. Luis. Fig. 14 surfaces126. This would provide an elegant way of lo- recording changes in the magnetic moment µ of the sam- cating magnetic systems close enough to the sensor and ple coupled as a change of magnetic flux to the SQUID would also enable in-situ reference measurements. How- [Fig. 4(b)]. Usually, M(Bext) is hysteretic, due to an en- ever, their use for the investigation of magnetic molecules ergy barrier created by magnetic anisotropy. Such hys- or nanoparticles arranged on surfaces is still awaiting. teresis loops reveal information on the reversal mechanisms, e.g. domain wall nucleation and propagation or the formation of topological magnetic states like vortices, co- 3. Techniques based on chemical functionalization herent rotation, or quantum tunneling of magnetization. Depending on the particle’s anisotropy, this requires the The above mentioned techniques can be further im- application of relatively large Bext, a difficult task when proved by chemically functionalizing the sensor’s surface dealing with superconducting materials. Measurements or the MNPs or both of them127. This usually pro- are usually done by careful alignment of Bext with respect vides high quality monolayers or even individual mag- to the nanoSQUID, to minimize the magnetic flux coupled to the loop and the JJs by B directly. The maxi- netic molecules at specific positions. For instance, Mn12 ext SMMs could be successfully grafted on Au, the preferred mum Bext will be limited by the upper critical field of the substrate for chemical binding, by introducing thiol superconducting material, e.g. ∼ 1 T for Nb films, unless 128 ultrathin films are used, which however increases signifi- groups in the clusters . In a further step, such Mn12 molecules could be individually isolated by a combination cantly Lk and hence the flux noise (see section III B 3). 129 of molecule and Au substrate functionalization . The vastest amount of dc magnetization studies per- This technique has also been applied to the de- formed on individual MNPs was provided by the pioneer- position of ferritin-based MNPs onto Au-shunted ing work of Wernsdorfer and co-workers. They were able 130 2 nanoSQUIDs . For this purpose, a 200 × 200 nm to measure magnetization curves of a number of MNPs window was opened through e-beam lithography onto a made of Ni, Co, TbFe3 and Co81Zr9Mo8Ni2 with sizes PMMA layer deposited on top of the nanoSQUID. This down to 100 × 50 × 8 nm2. Furthermore, they succeeded window was then covered with organic linkers that were in measuring the dc magnetization of the smallest MNPs later used to attach the ferritin MNPs. The success of ever detected to date. These are 3 nm diameter crys- this process was finally determined by AFM, showing ev- 3 talline Co MNPs (10 µB each) directly embedded into idence that few proteins were attached. the Nb film forming the nanoSQUID41. The detected magnetization switching process was attributed to an individual MNP located precisely at the cJJ, where the B. Magnetization measurements coupling factor is maximized. These studies also enabled the determination of the 2nd and 4th order anisotropy NanoSQUIDs can be applied to study the reversal terms in the magnetic anisotropy of the Co MNPs. Ad- of magnetization M of MNPs placed nearby. For this ditionally, many exciting phenomena were studied with purpose an external magnetic field Bext is swept while this technique. These include, e.g., the observation of 16

(a) (b)

Si 0 Φ Laser cantilever / beam NN Ni Φ tube z Ib Ib (c) y

Imod x FIG. 17. (a) Sketch of combined torque and

(Hz) nanoSQUID magnetometry on a Ni nanotube.

JJs f Imod ∆ (b,c) Simultaneously measured hysteresis loops (b) ΦNN(H), (c) ∆f(H). Arrows indicate H sweep direction. Dashed lines indicate discontinuities appearing in both ΦNN(H) and ∆f(H). 51 124 µ0H (mT) [after and ]

Stoner-Wohlfarth and Néel-Brown type of thermally as- magnetic native-oxide layer on the nanotube modifies the Fig.sisted 15 magnetization reversal in individual Co clusters magnetization reversal process at low temperatures125. 6 111 (25 nm, 10 µB) or the observation of macroscopic YBCO nanoSQUIDs were used for the investigation quantum tunneling of magnetization in BaFeCoTiO sin- of magnetization reversal in a Fe nanowire grown in- 5 131 53 gle particles (10 − 20 nm, 10 µB) . Magnetization re- side a CNT attached on top of the SQUID [Fig.18(a)]. versal triggered by rf field pulses on a 20 nm diameter Co Magnetization measurements were performed at 4.2 K NP was also reported132 and, recently, the effects of the in FLL by continuously sweeping H in the plane of antiferromagnetic-ferromagnetic exchange bias between a the SQUID loop, along the Fe wire axis. Rectangu- Co nanocluster and a CoO layer were revealed133. Micro- lar shaped hysteresis loops [Fig.18(b)] indicate a single metric SMM crystals were also investigated with an array domain state for the nanowire. The magnitude of the containing four microSQUIDs134. These experiments al- switching field suggests that magnetization reversal takes lowed observing the modulation of the small (10−7 K) place non-uniformly, e.g., by curling. These results agree tunnel splitting in Fe8 molecular clusters under the ap- very well with previous measurements on an individual plication of a transverse magnetic field135. nanowire using a micro-Hall bar18, albeit with a significantly improved signal-to-noise ratio. Similarly, YBCO Magnetization reversal mechanisms in single Ni nanoSQUIDs were used to detect the magnetization re- and permalloy nanotubes were investigated using versal of individual Co MNPs with magnetic moments Nb/HfTi/Nb-based nanoSQUIDs51,124,125. Experiments (1 − 30) × 106 µ at different temperatures ranging from were performed at 4.2 K with B = µ H applied along B ext 0 300 mK up to 80 K. These studies allowed the identi- the nanotube axis (z-axis), with the SQUID loop in the fication of two different reversal mechanisms which de- x-z plane. The nanoSQUID was mounted on an x-y-z pend on the dimensions and shape of the Co particles. stage below the bottom end of the nanotube which is The different reversal mechanisms are linked to the sta- affixed to an ultrasoft Si cantilever [Fig.17(a)]. The nan- bilization two different magnetic states, i.e., the (quasi) otube was positioned to maximize the flux Φ coupled NN single-domain and the vortex state115. to the nanoSQUID. While recording the SQUID output operated in FLL, simultaneously the magnetic torque ex- erted on the nanotube was detected, by recording the frequency shift ∆f on the cantilever resonance frequency C. Susceptibility measurements as a function of H. Measurements on a Ni nanotube showed discontinuities at the same values of H that were Even more demanding, nanoSQUIDs can also be used ascribed to switching of the magnetization along the nan- to quantify the response of a MNP to an oscillating mag- otube [Fig. 17(b)]. These experiments provided, on the netic field Bac = B0 cos(ωt), i.e., its frequency-dependent one hand, the magnetic field stray produced by the nan- magnetic susceptibility χac = χre +iχim, where χre is the otube’s end and, on the other, the volume magnetization, part going in-phase with Bac and χim is the out-of-phase giving evidence for the formation of a magnetic vortex- part. These quantities bear much information on the dy- like configuration in the nanotube. Measurements on an namic behavior of spins and the relaxation processes to individual permalloy nanotube evidenced the nucleation thermal equilibrium, the interaction between spins, and of magnetic vortices at the nanotube’s end before prop- the ensuing magnetic phase transitions. These measure- agating through its whole length, leading to the com- ments can be performed using SQUID-based susceptome- plete switching of the magnetization. Furthermore, it ters, usually in a gradiometric design to be insensitive to has been shown that a thin exchange-coupled antiferro- homogeneous external magnetic fields, but sensitive to 17

the imbalance produced by a MNP located in one of the (a) Monolayer dots Bulk-like coils [Fig. 4(c,d)]. χre and χim are directly accessible by 80 f (kHz) applying a homogeneous Bac via on-chip excitation coils 0.021 2.1 /kOe) and lock-in detecting the nanoSQUID output. Alterna- √ B 40 79 tively, SΦ can be measured, as it is directly related to µ 136 ( re χim through the fluctuation-dissipation theorem . The χ detection of χac demands high sensitivity, as the net os- 0 cillating polarization induced in the sample is, by far, 0.0 0.2 0.4 0.0 0.2 0.4 smaller than the total saturation magnetization. At best, T (K) T (K) broad-band frequency measurements must be performed (b) which also provide an easy way to filter out the 1/f noise χ T (mK) re 15 48 78 125 of the SQUIDs, therefore improving the effective sensi- 184 230 289 401 tivity of the sensor. Frequencies are usually restricted

(a.u.) χ to ∼ 1 MHz, mainly limited by the room temperature im χ amplifiers and the FLL circuit. One of the most controversial observations of quantum coherence in nanoscopic magnets was realized us- 10-2 100 102 104 10-2 100 102 104 ing the SQUID-based microsusceptometer developed by f (Hz) f (Hz) Ketchen et al.9. This device allowed the detection of the magnetic susceptibility of small spin populations ofFig. 17FIG. 19. Magnetic susceptibility χ measured with SQUID- natural horse-spleen ferritin137. For a sample with just based microsusceptometers. (a) Ferritin monolayer dots and 4 bulk sample: χre(T ) obtained at three different frequencies. 4 × 10 proteins (∼ 200 µB/protein), a resonance√ peak in both the out-of-phase component of χ and S The superparamagnetic blocking of the susceptibility is visible ac Φ below 50 mK in both cases [after Mart´ınez-Pérez et al.121]. has been observed and was attributed to the zero-field 137,138 (b) HoW10 SMM crystal: χre(f) (left) and χim(f) (right) splitting energy . This is the energy separating the measured at different T . two non-degenerated low-energy quantum states, i.e., the (anti-)symmetric combination of the classical states cor- responding to magnetization pointing (down) up. This posed of antiferromagnetic CoO leading to a tiny mag- interpretation and the magnitude of this zero-field split- netic moment of ∼ 10 µB per protein. Monolayer ar- ting (900 kHz) is still an object of debate. rangements of ferritin MNPs (total amount ∼ 107 pro- MNPs artificially grown inside ferritin were also stud- teins) were deposited by DPN directly onto the SQUID, ied using a SQUID-based microsusceptometer121. The maximizing the coupling between the samples and the 120 magnetic core with diameter of just a few nm was com- sensor’s pickup coils (see section IV A 2). Using Bac ∼ 0.1 mT, these experiments showed that ferritin-based MNPs arranged on surfaces retain their properties, still exhibiting superparamagnetic blocking of the magnetic FIG. 18. susceptibility [Fig. 19(a)]. Furthermore, these results (a) SEM image (a) SQUID loop allowed to determine experimentally the spin sensitiv- of Fe nanowire ity. This was done by determining the coupling, i.e., the Fe nanowire encapsulated measured flux signal coupled to the microsusceptome- in a CNT on top of a YBCO ter divided by the total magnetic moment of the parti- nanoSQUID. cle, which was located at an optimum position on top (b) Hysteresis of the field coil or close to the edge of the pickup-loop. loop Φ(H) of Together with the measured flux noise of the SQUID, 1/2 √ the Fe nanowire, this yielded Sµ ∼ 300 µB/ Hz. Additionally, a large detected by the amount of measurements on SMM micron-sized crystals 300 nm CNT SQUID. Left or powder at very low T were reported [Fig. 19(b)]. The axis corresponds large bandwidth of these susceptometers (1 mHz–1 MHz) 2 100 to magnetiza- (b) M enabled, e.g., the investigation of the relationship be- s tion signal M; tween quantum tunneling and spin-phonon interaction 1 50 the literature value for the and to point out novel and reliable molecular candidates

T=4.2 K ) A/m) 0 for quantum computing and low-temperature magnetic 6 saturation mag- 0 0 Φ 34,139–141 netization Ms = refrigerants (e.g., Refs. ). (m Φ M (10 1710 kA/m of Microsusceptometers were also used to detect the ac -1 -50 7 Fe is indicated magnetic susceptibility of just ∼ 9 × 10 Mn12 SMMs

-Ms as dashed lines. arranged as dot-like features containing 3–5 molecu- -2 -100 122 -150 -100 -50 0 50 100 150 [after Schwarz et lar layers . Measurements showed an evident de- µ H (mT) al.53] 0 crease of the magnetic relaxation time compared to that

Fig. 18 18 observed in crystalline Mn12. This phenomenon was this allows to integrate field coils around the pickup loop attributed to structural modifications of the surface- for susceptibility measurements and modulation coils in- arranged molecules leading to an effective decrease of ductively coupled to the SQUID loop for separate flux their activation energy. These sensors have been also modulation of the SQUID, i.e. without disturbing the applied to the investigation of quantum spin dynamics of signals to be detected by the pickup loop. The Si sub- 142 Fe4 SMMs grafted onto graphene flakes . strate is polished to form a corner, typically at a distance dcorner of a few tens of µm away from the center of the pickup loop. SQUID microscopes based on such sensors V. NANOSQUIDS FOR SCANNING SQUID use a mechanical lever for scanning. The SQUID chip is MICROSCOPY mounted on a cantilever with a small inclination angle α to the plane of the sample. The vertical pickup-loop to 144 In scanning SQUID microscopy (SSM) the high sensi- sample distance is then given by dcorner sin α . If the tivity of SQUIDs to magnetic flux is combined with high SQUID is well thermally linked to the liquid He bath for spatial resolution by scanning a sample under investi- operation at 4.2 K, the sample mounted in vacuum can 148 gation relative to a miniaturized SQUID sensor, or vice be heated to above ∼ 100 K . versa. A variety of SSM systems has been developed in The most important application of the IBM micro- the 1990s and refined since then. Those were based on scope was the pioneering work on the order parameter both, metallic low-Tc and high-Tc cuprate superconduc- symmetry of cuprate superconductors. Just to men- tors, although the majority of work focused on the low-Tc tion a few examples, this includes key experiments for devices. For a review on the developments of SSM in the providing clear evidence of dx2−y2 -wave pairing in the 1990s see Ref.143. cuprates by imaging fractional vortices along YBCO Obviously, miniaturized SQUID structures can signifi- GBJs149, the formation of half-integer flux quanta in cantly improve the spatial resolution and sensitivity to lo- cuprate tricrystals150 and in Nb/cuprate hybrid Joseph- cal magnetic field sources. A key issue is the requirement son junctions, forming zigzag-type JJs or huge arrays of to approach the surface of the samples under investiga- π-rings151. For more applications, see the review13. tion to a distance which is of the order of or even smaller Very similar devices, also based on the Nb multilayer than the SQUID size or pickup loop, in order to gain in technology, have been developed and used for SSM by spatial resolution by shrinking the lateral dimensions of the Stanford group of Moler and co-workers58,59. Based the structures. Several strategies for improving the spa- on the original microsusceptometer design of Ketchen tial resolution in SSM have been followed, which can be et al.44, these devices contain two oppositely wound divided into three approaches. The two conventional ap- pickup coils, to cancel homogeneous applied fields.√ Sen- proaches, developed in the 1990s use SQUID structures sors with√∼ 4 µm pickup-loop diameter√ achieved SΦ = 59,152 on planar substrates. One is based on the sensing of local 0.8 µΦ0/ Hz at 4 K and 0.25 µΦ0/ Hz below 0.5 K . fields by a miniaturized pickup loop, coupled to a SQUID The sensor’s substrate was cut by polishing, leading to sensor; the other is based on using miniaturized SQUID dcorner ∼ 25 µ. A capacitive approach control was used loops to which local magnetic signals are coupled directly to monitor the probe-to-sample distance. These micro- (section V A). A very recently developed third approach susceptometers were largely improved by using a terraced uses the SQUID-on-tip (SOT), i.e. a SQUID deposited cantilever obtained through a multilayer lithography pro- directly on top of a nanotip (section V B). cess. In this way the pickup loop stands above the rest of the structure lying at just 300 nm above the sample surface. Additionally the pickup loop diameters were A. SQUID microscopes using devices on planar reduced down to 600 nm using focused ion beam (FIB) substrates milling152. Based on these SQUID sensors, the Stanford group has developed a SQUID microscope operating at 153 SQUID microscopes developed at IMB research by temperatures down to 20 mK in a dilution refrigerator . 144 Kirtley et al. are based on Nb/Al-AlOx/Nb technol- The SSM system of the Stanford group has been very ogy. The sensors are based on a single SQUID loop successfully applied to a variety of interesting systems. with an integrated pickup loop44. The pickup loops have Just to give a few examples, this includes the study of diameters down to ∼ 4 µm and are connected via well edge currents in topological insulators154, surface mag- 155 156 shielded superconducting thin film leads to the SQUID netic states and twin walls at the LaAlO3/SrTiO3 loop at typically ∼ 1 mm distance on the same chip145. interface, or unpaired spins in metals33. This technology has also been used to realize a miniature As an alternative approach, the group of Hasselbach vector magnetometer for SSM by using three SQUIDs and co-workers at Institut Néel, Grenoble developed with orthogonal pickup loops on a single chip146. As an SSM based on miniaturized Nb and Al SQUIDs a key advantage, the IBM designs are based on the loops with constriction JJs71, very similar to the ones very mature Nb multilayer SIS technology, including pat- of the Wernsdorfer group38. This approach allows for terning by photolithography, that allows e.g. using the a relatively simple single layer fabrication process with HYPRES147 process for sensor fabrication. Moreover, prospects of strong miniaturization. To achieve at the 19 same time small probe-to-sample distances, the sensor’s substrate was cut using a dicing machine and a mesa was defined by means of reactive ion etching so that the distance between the SQUID and apex of the mesa (‘tip’) was only 2 − 3 µm. With an inclination angle α ∼ 5◦, this gives a smallest vertical distance to a sample surface of ∼ 0.26 µm. The SSM setup is combined with force microscopy, based on the use of a mechanically excited quartz tuning fork and operates in a dilution refrigerator, achieving minimum SQUID and sample temperatures of 0.45 K73. Very recently, in a modified setup, a SQUID- to-sample distance of 420 nm has been demonstrated, in a setup with 40 mK base temperature157. The SSM system of the Grenoble group has been applied to the investigation of basic properties of superconductors. This includes, e.g., studies on the direct observation of the localized superconducting state around holesFig. 19.9FIG. 20. SQUID-on-tip (SOT): (a) schematic of a sharp in perforated Al films158 or on the Meissner-Ochsenfeld quartz pipette with superconducting leads, connecting to the effect and absence of the Meissner state in the ferromag- SOT at the bottom end; inset shows magnified view. (b) 159 SEM image of a Nb SOT having a diameter of 238 nm. netic superconductor UCoGe . Reprinted by permission from Macmillan Publishers Ltd: Na- ture Nanotechnology92, copyright (2013).

B. SQUID-on-tip (SOT) microscope reaches values below 10 pH, dominated by the kinetic An important breakthrough in the field of inductance of the thin superconducting layer. Although nanoSQUIDs applied to SSM was achieved recently these nanoSQUIDs exhibit hysteretic IVCs, operation with the implementation of the SQUID-on-tip (SOT) with voltage bias and reading out the resulting current by the Zeldov group at the Weizman Institute of signal with an SSA enables the detection of the intrinsic Science91,92. This device is based on the deposition of flux noise of the devices. The SOTs can be operated in a nanoSQUID directly on the apex of a sharp quartz large magnetic fields up to ∼ 1 T (limited by the upper pipette [Fig.20]. The fact that the nanoSQUID is critical fields of the superconducting materials). So far, located on a sharp tip reduces the possible minimum flux biasing to maintain the optimum working point probe-to-sample distances to below 100 nm, boosting during continuous external field sweep is not possible. enormously the spatial resolution of the microscope. Al, By adjusting the external magnetic field to values which Nb and Pb nanoSQUIDs based on Dayem bridges are yield large transfer functions, these devices exhibit√ shadow-evaporated in a three-angle process, without extraordinary low flux noise levels down to 50 nΦ0/ Hz requiring any lithography or milling steps. For this for the Pb SOTs92. The latter varies, depending on purpose, a quartz pipette is first pulled to form a sharp the biasing external magnetic field. For a magnetic hollow tip with 40 − 300 nm inner diameter. By means dipole located at the center of the loop with orientation of a laser diode parallel to the tip, the latter is aligned perpendicular to the loop plane (assuming an infinitely pointing down towards the source which defines the narrow width of the loop, i.e. the approximation used by ◦ 44 0 position. Then a thin layer (< 10 nm) of supercon- Ketchen et√ al. ), this translates into a spin sensitivity ducting material is deposited, followed by two thicker of 0.38 µB/ Hz, i.e. the best spin sensitivity reported so leads (> 25 nm) deposited at ±100◦. The resulting far for a nanoSQUID. weak links formed at the tip apex between these two A device able of distinguishing in-plane and out-of- leads constitute two Dayem bridges. Special care must plane magnetic signals was also reported160. This is be taken for fabricating the Nb and Pb sensors. The achieved by using a pipette with θ-shaped cross section to former require the previous deposition of a thin AlOx form a three JJ SQUID (3JSOT). This tip is later milled buffer layer to prevent contamination from the quartz by FIB leading to a V shaped apex with two oblique tip. A dedicated ultra-high vacuum e-beam evaporation nanoloops connected in parallel. By measuring the de- system was used for depositing Nb from a point source. pendence of the maximum critical current on the exter- On the other hand, the so far most sensitive Pb sensors nally applied in-plane and out-of-plane magnetic fields require the use of a He cooling system for the tips during Ic(H||,H⊥), it is possible to determine all the geometrical deposition to prevent the formation of islands due to the and electric parameters of the device. Field components large surface mobility of these atoms at higher tempera- can be decoupled by biasing the 3JSOT at specific fields tures. This procedure lead to the smallest nanoSQUIDs (H||,H⊥) in which Ic depends strongly on one of the two fabricated so far, with effective nanoloop diameters orthogonal components of the magnetic field while being down to 50 nm. The resulting inductance of the loop insensitive to the other. As a drawback, this device is 20 not able of distinguishing both in-plane and out-of-plane structures, they can offer significantly reduced√ flux noise, components of the magnetic flux simultaneously, but only down to the level of a few tens of nΦ0/ √Hz, cor- when operated at different flux biasing points. responding to spin sensitivities around 1µB/ Hz and For SSM, a system operating in a 3He system with improved spatial resolution for scanning SQUID mi- 300 mK base temperature has been developed, with the croscopy. Hence, strongly miniaturized SQUIDs are very SOT glued on a quartz tuning fork, to operate the system promising detectors for investigating tiny and strongly also in an magnetic force microscopy mode. This allows localized magnetic signals produced, e.g., by magnetic scanning (using piezo-scanners) at extremely small tip- nanoparticles or for high-resolution scanning SQUID mi- to-sample distances of only a few nm161. A spatial resolu- croscopy. Very recent advances, including the demon- tion below 120 nm was demonstrated by imaging vortices stration of single spin sensitivity and a breakthrough in in Nb thin films with a 117 nm-diameter Pb SOT92. spatial resolution of scanning SQUID microscopy open The SOT-SSM system has been successfully applied up promising perspectives for applications in nanoscale to the study of vortex trajectories in superconducting magnetism of condensed matter systems. thin films, allowing the investigation of the influence of the pining force landscape162. More recently, this tool was used to observe nanoscopic magnetic structures such as ferromagnetic metallic nanoislands at the ACKNOWLEDGMENTS 163 LaMnO3/SrTiO3 interface or magnetic nanodomains in magnetic topological insulators164. We gratefully acknowledge valuable contributions by the nanoSQUID team at Tübingen, M. Kemmler, J. Nagel, R. Wölbing,T. Schwarz, B. Müller,S. Hess VI. SUMMARY AND OUTLOOK and R. Kleiner, and by our collaborators O. Kieler and A. Zorin et al. at PTB Braunschweig, T. Schurig et al. at Significant progress in thin film fabrication and patter- PTB Berlin, B. Büchner et al. at IFW Dresden, M. Pog- ing technologies has enabled the development of strongly gio et al. at Univ. Basel, D. Grundler et al. at TU miniaturized dc SQUIDs with loop sizes on the microm- Munich and EPFL Lausanne, A. Fontcuberta i Morral eter scale (microSQUIDs) or even with sub-micrometer at EPFL Lausanne and J. Seséat INA Zaragoza. This dimensions (nanoSQUIDs), or SQUIDs coupled to minia- work was funded by the Alexander von Humboldt Foun- turized pickup loops. Such devices are based on a vari- dation, the Nachwuchswissenschaftlerprogramm of the ety of Josephson junctions, intersecting the SQUID loop, Univ. Tübingen,the Deutsche Forschungsgemeinschaft, many of them also on the sub-micrometer scale. As via projects KO 1303/13-1 and SFB/TRR21, and by the a key advantage of such strongly miniaturized SQUID EU-FP6-COST Action MP1201.

1 J. Clarke and A. I. Braginski, eds., The SQUID Hand- 9 M. B. Ketchen, T. Kopley, and H. Ling, “Miniature book, Vol. I: Fundamentals and Technology of SQUIDs SQUID susceptometer,” Appl. Phys. Lett. 44, 1008–1010 and SQUID systems (Wiley-VCH, Weinheim, 2004). (1984). 2 R. Kleiner, D. Koelle, F. Ludwig, and J. Clarke, “Super- 10 F. P. Rogers, A device for experimental observation of conducting QUantum interference devices: State-of-the- flux vortices trapped in superconducting thin films, Ph.D. art and applications,” Proc. IEEE 92, 1534–1548 (2004). thesis, MIT, Cambridge, MA (1983). 3 J. Clarke and A. I. Braginski, eds., The SQUID Hand- 11 W. Wernsdorfer, K. Hasselbach, D. Mailly, B. Barbara, book, Vol. II: Applications of SQUIDs and SQUID Sys- A. Benoit, L. Thomas, and G. Suran, “DC-SQUID mag- tems (Wiley-VCH, Weinheim, 2006). netization measurements of single magnetic particles,” J. 4 M. Gurvitch, M. A. Washington, and H. A. Huggins, Magn. Magn. Mat. 145, 33–39 (1995). “High quality refractory Josephson tunnel junctions uti- 12 W. Wernsdorfer, “Classical and quantum magnetization lizing thin aluminum layers,” Appl. Phys. Lett. 42, 472– reversal studied in nanometersized particles and clus- 474 (1983). ters,” Adv. Chem. Phys. 118, 99–190 (2001), arXiv:cond- 5 Richard F. Voss, Robert B. Laibowitz, and Alec N. mat/0101104v1. Broers, “Niobium nanobridge dc SQUID,” Appl. Phys. 13 J. R. Kirtley, “Fundamental studies of superconductors Lett. 37, 656–658 (1980). using scanning magnetic imaging,” Rep. Prog. Phys. 73, 6 A. N. Broers, W. W. Molzen, J. J. Cuomo, and N. D. 126501 (2010). Wittels, “Electron-beam fabrication of 80 A˚ metal struc- 14 C. Granata and A. Vettoliere, “Nano superconduct- tures,” Appl. Phys. Lett. 29, 596–598 (1976). ing quantum interference device: A powerful tool for 7 C. D. Tesche and J. Clarke, “DC SQUID: Noise and op- nanoscale investigations,” Phys. Rep. 614, 1–69 (2016), timization,” J. Low Temp. Phys. 29, 301–331 (1977). arXiv:1505.06887 [cond-mat.supr-con]. 8 Roger H. Koch, D. J. Van Harlingen, and John Clarke, 15 Eike Schäfer-Nolte,Lukas Schlipf, Markus Ternes, Friede- “Quantum noise theory for the dc SQUID,” Appl. Phys. mann Reinhard, Klaus Kern, and Jörg Wrachtrup, Lett 38, 380–382 (1981). “Tracking temperature-dependent relaxation times of fer- 21

ritin nanomagnets with a wideband quantum spectrome- (2008). ter,” Phys. Rev. Lett. 113, 217204 (2014). 33 Hendrik Bluhm, Julie A. Bert, Nicholas C. Koshnick, 16 L. Thiel, D. Rohner, M. Ganzhorn, P.Appel, E. Neu, Martin E. Huber, and Kathryn A. Moler, “Spinlike sus- B. Müller,R. Kleiner, D. Koelle, and P. Maletinsky, ceptibility of metallic and insulating thin films at low tem- “Quantitative nanoscale vortex imaging using a perature,” Phys. Rev. Lett. 103, 026805 (2009). cryogenic quantum magnetometer,” Nature Nan- 34 M. J. Mart´ınez-Pérez, J. Sesé, F. Luis, R. Córdoba, otechnology (2016), 10.1038/nnano.2016.63, dOI: D. Drung, T. Schurig, E. Bellido, R. de Miguel, C. Gómez- 10.1038/nnano.2016.63, arXiv:1511.02873 [cond-mat.mes- Moreno, A. Lostao, and D. Ru´ız-Molina, “Ultrasensi- hall]. tive broad band SQUID microsusceptometer for magnetic 17 Marc Ganzhorn, Svetlana Klyatskaya, Mario Ruben, measurements at very low temperatures,” IEEE Trans. and Wolfgang Wernsdorfer, “Carbon nanotube nano- Appl. Supercond. 21, 345–348 (2011). electromechanical systems as magnetometers for single- 35 M. J. Mart´ınez-Pérez, J. Sesé,F. Luis, D. Drung, and molecule magnets,” ACS Nano 7, 6225–6236 (2013). T. Schurig, “Note: Highly sensitive superconducting 18 K. Lipert, S. Bahr, F. Wolny, P. Atkinson, U. Weißker, quantum interference device microsusceptometers operat- T. Mühl, O. G. Schmidt, B. Büchner, and R. Klin- ing at high frequencies and very low temperatures inside geler, “An individual iron nanowire-filled carbon nan- the mixing chamber of a dilution refrigerator,” Rev. Sci. otube probed by micro-Hall magnetometry,” Appl. Phys. Instr. 81, 016108 (2010). Lett. 97, 212503 (2010). 36 F. C. Wellstood, C. Urbina, and John Clarke, “Hot- 19 The term nanoSQUID denotes strongly miniaturized thin electron effects in metals,” Phys. Rev. B 49, 5942–5955 film SQUIDs with lateral dimensions in the submicrome- (1994). ter range. However, some devices described here and also 37 D. Drung and M. Mück, “SQUID Electronics,” in The various statements made also refer to slightly larger struc- SQUD Handbook, Vol. I: Fundamentals and Technology of tures, which sometimes are denoted as microSQUIDs. SQUIDs and SQUID Systems, edited by John Clarke and Throughout the text, we do not make this discrimination. Alex I. Braginski (Wiley-VCH, Weinheim, 2004) Chap. 4, 20 F. London, Superfluids (Wiley, New York, 1950). pp. 127–170. 21 B. D. Josephson, “Possible new effects in superconductive 38 W. Wernsdorfer, “From micro- to nano-SQUIDs: appli- tunneling,” Phys. Lett. 1, 251–253 (1962). cations to nanomagnetism,” Supercond. Sci. Technol. 22, 22 P. W. Anderson and J. M. Rowell, “Probable observation 064013 (2009). of the Josephson superconducting tunneling effect,” Phys. 39 R. Russo, C. Granata, E. Esposito, D. Peddis, C. Can- Rev. Lett. 10, 230–232 (1963). nas, and A. Vettoliere, “Nanoparticle magnetization mea- 23 K. K. Likharev, “Superconducting weak links,” Rev. Mod. surements by a high sensitive nano-superconducting quan- Phys. 51, 101–159 (1979). tum interference device,” Appl. Phys. Lett. 101, 122601 24 W. C. Stewart, “Current-voltage characteristics of (2012). Josephson junctions,” Appl. Phys. Lett 12, 277–280 40 Carmine Granata, Roberto Russo, Emanuela Esposito, (1968). Antonio Vettoliere, Maurizio Russo, Anna Musinu, Da- 25 D.E. McCumber, “Effect of ac impedance of dc voltage- vide Peddis, and Dino Fiorani, “Magnetic properties of current characteristics of Josephson junctions,” J. Appl. iron oxide nanoparticles investigated by nanoSQUIDs,” Phys. 39, 3113–3118 (1968). Eur. Phys. J. B 86, 272 (2013). 26 B. Chesca, R. Kleiner, and D. Koelle, “SQUID The- 41 M. Jamet, W. Wernsdorfer, C. Thirion, D. Mailly, ory,” in The SQUD Handbook, Vol. I: Fundamentals and V. Dupuis, P. Mélinon, and A. Pérez, “Magnetic Technology of SQUIDs and SQUID systems, edited by anisotropy of a single cobalt nanocluster,” Phys. Rev. John Clarke and Alex I. Braginski (Wiley-VCH, Wein- Lett. 86, 4676–4679 (2001). heim, 2004) Chap. 2, pp. 29–92. 42 M. Hatridge, R. Vijay, D. H. Slichter, John Clarke, and 27 R. C. Jaklevic, J. Lambe, A. H. Silver, and J. E. Mer- I. Siddiqi, “Dispersive magnetometry with a quantum cereau, “Quantum interference effects in Josephson tun- limited SQUID parametric amplifier,” Phys. Rev. B 83, neling,” Phys. Rev. Lett. 12, 159–160 (1964). 134501 (2011). 28 R. Kleiner and D. Koelle, “Basic Properties of Supercon- 43 E. M. Levenson-Falk, R. Vijay, N. Antler, and I. Sid- ductivity,” in The SQUD Handbook, Vol. I: Fundamentals diqi, “A dispersive nanoSQUID magnetometer for ultra- and Technology of SQUIDs and SQUID systems, edited low noise, high bandwidth flux detection,” Supercond. Sci. by John Clarke and Alex I. Braginski (Wiley-VCH, Wein- Technol. 26, 055015 (2013). heim, 2004) Chap. Appendix 1, pp. 357–366. 44 M. Ketchen, D. Awschalom, W. Gallagher, A. Klein- 29 I I Soloviev, N V Klenov, A E Schegolev, S V Bakurskiy, sasser, R. Sandstrom, J. Rozen, and B. Bumble, “De- and M Yu Kupriyanov, “Analytical derivation of dc squid sign, fabrication, and performance of integrated miniature response,” Supercond. Sci. Technol. 29, 094005 (2016). SQUID susceptometers,” IEEE Trans. Magn. 25, 1212– 30 D. Koelle, R. Kleiner, F. Ludwig, E. Dantsker, and 1215 (1989). 45 44 John Clarke, “High-transition-temperature superconduct- φµ = 2π/a in cgs units, as derived by Ketchen et al. . 44 ing quantum interference devices,” Rev. Mod. Phys. 71, The spin sensitivity Sn in relates to our definition as p 631–686 (1999). Sn = Sµ/µB, i.e. Sn has the units of number of spins 31 √ Roger H. Koch, David P. DiVincenzo, and John Clarke, (of moment µB) per Hz. “Model for 1/f flux noise in SQUIDs and qubits,” Phys. 46 V. Bouchiat, “Detection of magnetic moments using a Rev. Lett. 98, 267003 (2007). nano-squid: limits of resolution and sensitivity in near- 32 S. Sendelbach, D. Hover, A. Kittel, M. Mück, John M. field squid magnetometry,” Supercond. Sci. Technol. 22, Martinis, and R. McDermott, “Magnetism in SQUIDs at 064002 (2009). millikelvin temperatures,” Phys. Rev. Lett. 100, 227006 22

47 David L. Tilbrook, “NanoSQUID sensitivity for isolated and Mark B. Ketchen, “Scanning superconducting quan- dipoles and small spin populations,” Supercond. Sci. tum interference device susceptometry,” Rev. Sci. Instr. Technol 22, 064003 (2009). 72, 2361–2364 (2001). 48 J. Nagel, K. B. Konovalenko, M. Kemmler, M. Tu- 59 Martin E. Huber, Nicholas C. Koshnick, Hendrik Bluhm, rad, R. Werner, E. Kleisz, S. Menzel, R. Klingeler, Leonard J. Archuleta, Tommy Azua, Per G. Björnsson, B. Büchner, R. Kleiner, and D. Koelle, “Resistively Brian W. Gardner, Sean T. Halloran, Erik A. Lucero, and shunted YBa2Cu3O7 grain boundary junctions and low- Kathryn A. Moler, “Gradiometric micro-squid suscep- noise SQUIDs patterned by a focused ion beam down to tometer for scanning measurements of mesoscopic sam- 80 nm linewidth,” Supercond. Sci. Technol. 24, 015015 ples,” Rev. Sci. Instrum. 79, 053704 (2008). (2011). 60 D. Drung, C. Aßmann, J. Beyer, A. Kirste, M. Peters, 49 The current J through an infinitely thin wire, forming F. Ruede, and Th. Schurig, “Highly sensitive and easy- a loop with radius a in the x − y plane and centered at to-use SQUID sensors,” IEEE Trans. Appl. Supercond. the origin, induces a field BJ = µ0J/(2a) · eˆz, at the 17, 699–702 (2007). center of the loop. Hence, for a magnetic dipole placed at 61 M. J. Mart´ınez-Pérez, J. Sesé, R. Córdoba, F. Luis, the origin rµ = 0 and pointing in z direction, eˆµ = eˆz, D. Drung, and T. Schurig, “Circuit edit of superconduct- Eq. (10) yields φµ = eˆz · BJ (rµ)/J = µ0/(2a), i.e. the ing microcircuits,” Supercond. Sci. Technol 22, 125020 same result as derived by Ketchen et al.44. (2009). 50 J. Nagel, O. F. Kieler, T. Weimann, R. Wölbing, 62 D. Drung, J.-H. Storm, F. Ruede, A. Kirste, M. Re- J. Kohlmann, A. B. Zorin, R. Kleiner, D. Koelle, and gin, T. Schurig, A. M. Repollés, J. Sesé, and M. Kemmler, “Superconducting quantum interference de- F. Luis, “Thin-film microsusceptometer with integrated vices with submicron Nb/HfTi/Nb junctions for investi- nanoloop,” IEEE Trans. Appl. Supercond. 24, 1600206 gation of small magnetic particles,” Appl. Phys. Lett. 99, (2014). 032506 (2011). 63 Thomas Schurig, “Making SQUIDs a practical tool for 51 J. Nagel, A. Buchter, F. Xue, O. F. Kieler, T. Weimann, quantum detection and material characterization in the J. Kohlmann, A. B. Zorin, D. Rüffer, E. Russo-Averchi, micro- and nanoscale,” J. Phys.: Conf. Ser. 568, 032015 R. Huber, P. Berberich, A. Fontcuberta i Morral, (2014). D. Grundler, R. Kleiner, D. Koelle, M. Poggio, and 64 M. Schmelz, R. Stolz, V. Zakosarenko, S. Anders, M. Kemmler, “Nanoscale multifunctional sensor formed L. Fritzsch, H. Roth, and H.-G. Meyer, “Highly sensitive by a Ni nanotube and a scanning Nb nanoSQUID,” Phys. miniature SQUID magnetometer fabricated with cross- Rev. B 88, 064425 (2013). type Josephson tunnel junctions,” Physica C 476, 77–80 52 T. Schwarz, J. Nagel, R. Wölbing, M. Kemmler, (2012). R. Kleiner, and D. Koelle, “Low-noise nano supercon- 65 M. Schmelz, Y. Matsui, R. Stolz, V. Zakosarenko, ducting quantum interference device operating in tesla T. Schönau,S. Anders, S. Linzen, H. Itozaki, and H.-G. magnetic fields,” ACS Nano 7, 844–850 (2013). Meyer, “Investigation of all niobium nano-SQUIDs based 53 T. Schwarz, R. Wölbing,C. F. Reiche, B. Müller,M. J. on sub-micrometer cross-type Josephson junctions,” Su- Mart´ınez-Pérez,T. Mühl,B. Büchner, R. Kleiner, and percond. Sci. Technol. 28, 015004 (2015). 66 D. Koelle, “Low-noise YBa2Cu3O7 nano-SQUIDs for per- D. Hagedorn, O. Kieler, R. Dolata, R. Behr, F. Müller, forming magnetization-reversal measurements on mag- J. Kohlmann, and J. Niemeyer, “Modified fabrica- netic nanoparticles,” Phys. Rev. Appl. 3, 044011 (2015). tion of planar sub-µm superconductor-normal metal- 54 R. Wölbing,J. Nagel, T. Schwarz, O. Kieler, T. Weimann, superconductor Josephson junctions for use in a Joseph- J. Kohlmann, A. B. Zorin, M. Kemmler, R. Kleiner, and son arbitrary waveform synthesizer,” Supercond. Sci. D. Koelle, “Nb nano superconducting quantum interfer- Technol. 19, 294–298 (2006). ence devices with high spin sensitivity for operation in 67 S. Bechstein, F. Ruede, D. Drung, J.-H. Storm, O. F. magnetic fields up to 0.5 T,” Appl Phys. Lett. 102, 192601 Kieler, J. Kohlmann, T. Weimann, and T. Schurig, (2013). “HfTi-nanoSQUID gradiometers with high linearity,” 55 R. Wölbing,T. Schwarz, B. Müller,J. Nagel, M. Kemm- Appl. Phys. Lett. 106, 072601 (2015). ler, R. Kleiner, and D. Koelle, “Optimizing the spin sensi- 68 M. J. Mart´ınez-Pérez, D. Gella, B. Müller, V. Mo- tivity of grain boundary junction nanoSQUIDs – towards rosh, R. Wölbing,J. Sesé,O. Kieler, R. Kleiner, and detection of small spin systems with single-spin resolu- D. Koelle, “Three-axis vector nano superconducting quan- tion,” Supercond. Sci. Technol. 27, 125007 (2014). tum interference device,” ACS Nano (2016), 10.1021/ac- 56 φµ depends significantly on the loop width, thickness d snano.6b02218, arXiv:1604.07195v2 [cond-mat.supr-con]. 69 and λL. For example for a dipole centered at a circu- Carmine Granata, Antonio Vettoliere, Roberto Russo, lar SQUID loop with inner radius a = 500 nm, outer Matteo Fretto, Natascia De Leo, and Vincenzo Lac- radius R = 2 µm, and d = λL = 100 nm one finds quaniti, “Three-dimensional spin nanosensor based on φµ = 3.5 nΦ0/µB, i.e. a factor 1.6 smaller φµ as obtained reliable tunnel Josephson nano-junctions for nanomag- 44 from Ref. (with R = a = 500 nm); φµ decreases further netism investigations,” Appl. Phys. Lett. 103, 102602 with decreasing ratio d/λL. (2013). 57 P. Josephs-Franks, L. Hao, A. Tzalenchuk, J. Davies, 70 P. W. Anderson and A. H. Dayem, “Radio-frequency ef- O. Kazakova, J. C. Gallop, L. Brown, and J. C. Mac- fects in superconducting thin film bridges,” Phys. Rev. farlane, “Measurement of the spatial sensitivity of minia- Lett. 13, 195–197 (1964). ture SQUIDs using magnetic-tipped STM,” Supercond. 71 K. Hasselbach, C. Veauvy, and D. Mailly, “MicroSQUID Sci. Technol. 16, 1570–1574 (2003). magnetometry and magnetic imaging,” Physica C 332, 58 Brian W. Gardner, Janice C. Wynn, Per G. Björnsson, 140–147 (2000). Eric W. J. Straver, Kathryn A. Moler, John R. Kirtley, 23

72 M. Faucher, T. Fournier, B. Pannetier, C. Thirion, 88 L. Hao, C. Aßmann, J. C. Gallop, D. Cox, F. Ruede, W. Wernsdorfer, J.C. Villegier, and V. Bouchiat, “Nio- O. Kazakova, P. Josephs-Franks, D. Drung, and Th. bium and niobium nitride SQUIDs based on anodized Schurig, “Detection of single magnetic nanobead with nanobridges made with an atomic force microscope,” a nano-superconducting quantum interference device,” Physica C 368, 211–217 (2002). Appl. Phys. Lett. 98, 092504 (2011). 73 C. Veauvy, K. Hasselbach, and D. Mailly, “Scanning mu- 89 Lei Chen, Wolfgang Wernsdorfer, Christos Lampropoulos, superconduction quantum interference device force micro- George Christou, and Irinel Chiorescu, “On-chip SQUID scope,” Rev. Sci. Instr. 73, 3825–3830 (2002). measurements in the presence of high magnetic fields,” 74 C. Granata, A. Vettoliere, R. Russo, E. Esposito, Nanotechnology 21, 405504 (2010). M. Russo, and B. Ruggiero, “Supercurrent decay in 90 Soumen Mandal, Tobias Bautze, Oliver A. Williams, nano-superconducting quantum interference devices for CécileNaud, Etienne´ Bustarret, Franck Omnès,Pierre intrinsic magnetic flux resolution,” Appl. Phys. Lett. 94, Rodière, Tristan Meunier, Christopher Bäuerle,and Lau- 062503 (2009). rent Saminadayar, “The diamond superconducting quan- 75 D. Hazra, J. R. Kirtley, and K. Hasselbach, “Nano- tum interference device,” ACS Nano 5, 7144–7148 (2011). superconducting quantum interference devices with con- 91 Amit Finkler, Yehonathan Segev, Yuri Myasoedov, tinuous read out at millikelvin temperatures,” Appl. Phys. Michael L. Rappaport, Lior Ne’eman, Denis Vasyukov, Lett. 103, 093109 (2013). Eli Zeldov, Martin E. Huber, Jens Martin, and Amir 76 Aico G. P. Troeman, Hendrie Derking, Bert Borger, Jo- Yacoby, “Self-aligned nanoscale SQUID on a tip,” Nano hannes Pleikies, Dick Veldhuis, and Hans Hilgenkamp, Lett. 10, 1046–1049 (2010). “NanoSQUIDs based on niobium constrictions,” Nano 92 Denis Vasyukov, Yonathan Anahory, Lior Embon, Dorri Lett. 7, 2152–2156 (2007). Halbertal, Jo Cuppens, Lior Ne’eman, Amit Finkler, 77 R. Vijay, J. D. Sau, Marvin L. Cohen, and I. Siddiqi, “Op- Yehonathan Segev, Yuri Myasoedov, Michael L. Rappa- timizing anharmonicity in nanoscale weak link Josephson port, Martin E. Huber, and Eli Zeldov, “A scanning junction oscillators,” Phys. Rev. Lett. 103, 087003 (2009). superconducting quantum interference device with single 78 V. Bouchiat, M. Faucher, C. Thirion, W. Wernsdorfer, electron spin sensitivity,” Nature Nanotechnol. 8, 639–644 T. Fournier, and B. Pannetier, “Josephson junctions and (2013). superconducting quantum interference devices made by 93 J.-P. Cleuziou, W. Wernsdorfer, V. Bouchiat, T. On- local oxidation of niobium ultrathin films,” Appl. Phys. dar¸cuhu, and M. Monthioux, “Carbon nanotube super- Lett. 79, 123–125 (2001). conducting quantum interference device,” Nature Nan- 79 R. Vijay, E. M. Levenson-Falk, D. H. Slichter, and I. Sid- otech. 1, 53–59 (2006). diqi, “Approaching ideal weak link behavior with three 94 C. Girit, V. Bouchiat, O. Naaman, Y. Zhang, M. F. Crom- dimensional aluminum nanobridges,” Appl. 96, 223112 mie, A. Zettl, and I. Siddiqi, “Tunable graphene dc su- (2010). perconducting quantum interference device,” Nano Lett. 80 N. Antler, E.M. Levenson-Falk, R. Naik, Y.-D. Sun, 9, 198–199 (2009). A. Narla, R. Vijay, and I. Siddiqi, “In-plane mag- 95 L. Angers, F. Chiodi, G. Montambaux, M. Ferrier, netic field tolerance of a dispersive aluminum nanobridge S. Guéron,H. Bouchiat, and J. C. Cuevas, “Proximity SQUID magnetometer,” Appl. Phys. Lett. 102, 232602 dc squids in the long-junction limit,” Phys. Rev. B 77, (2013). 165408 (2008). 81 D. Hazra, J. R. Kirtley, and K. Hasselbach, “Nano- 96 P Spathis, S Biswas, S Roddaro, L Sorba, F Giazotto, and superconducting quantum interference devices with sus- F Beltram, “Hybrid inas nanowire–vanadium proximity pended junctions,” Appl. Phys. Lett. 104, 152603 (2014). SQUID,” Nanotechnology 22, 105201 (2011). 82 S. K. H. Lam and D. L. Tilbrook, “Development of a nio- 97 Francesco Giazotto, Joonas T. Peltonen, Matthias bium nanosuperconducting quantum interference device Meschke, and Jukka P. Pekola, “Superconducting quan- for the detection of small spin populations,” Appl. Phys. tum interference proximity transistor,” Nature Phys. 6, Lett. 82, 1078–1080 (2003). 254–259 (2010). 83 S. K. H. Lam, “Noise properties of SQUIDs made 98 R. N. Jabdaraghi, M. Meschke, and J. P. Pekola, “Non- from nanobridges,” Supercond. Sci. Technol. 19, 963–967 hysteretic superconducting quantum interference proxim- (2006). ity transistor with enhanced responsivity,” Appl. Phys. 84 Simon K. H. Lam, John R. Clem, and Wenrong Yang, Lett. 104, 082601 (2014). “A nanoscale SQUID operating at high magnetic fields,” 99 Alberto Ronzani, Carles Altimiras, and Francesco Nanotechnology 22, 455501 (2011). Giazotto, “Highly sensitive superconducting quantum- 85 P. F. Vohralik and S. K. H. Lam, “NanoSQUID detection interference proximity transistor,” Phys. Rev. Applied 2, of magnetization from ferritin nanoparticles,” Supercond. 024005 (2014). Sci. Technol. 22, 064007 (2009). 100 Mohammad Alidoust, Klaus Halterman, and Jacob Lin- 86 L. Hao, J. C. Macfarlane, J. C. Gallop, D. Cox, J. Beyer, der, “Singlet-triplet superconducting quantum magne- D. Drung, and T. Schurig, “Measurement and noise per- tometer,” Phys. Rev. B 88, 075435 (2013). formance of nano-superconducting-quantum-interference 101 R. Arpaia, M. Arzeo, S. Nawaz, S. Charpentier, F. Lom- devices fabricated by focused ion beam,” Appl. Phys. bardi, and T. Bauch, “Ultra low noise YBa2Cu3O7−δ Lett. 92, 192507 (2008). nano superconducting quantum interference devices im- 87 A. Blois, S. Rozhko, L. Hao, J. C. Gallop, and E. J. plementing nanowires,” App. Phys. Lett. 104, 072603 Romans, “Proximity effect bilayer nano superconducting (2014). quantum interference devices for millikelvin magnetome- 102 H. Hilgenkamp and J. Mannhart, “Grain boundaries in try,” J. Appl. Phys. 114, 233907 (2013). high-Tc superconductors,” Rev. Mod. Phys. 74, 485–549 (2002). 24

103 Francesco Tafuri and John R. Kirtley, “Weak links in high “Manipulation of single magnetic protein particles using critical temperature superconductors,” Rep. Prog. Phys. atomic force microscopy,” J. Magn. Magn. Mater. 272– 68, 2573–2663 (2005). 276, E1231–E1233 (2004). 104 Francesco Tafuri, Davide Massarotti, Luca Galletti, 118 D. Gella, Master’s thesis, University of Zaragoza (2015). Daniela Stornaiuolo, Domenico Montemurro, Luigi Lon- 119 Richard D. Piner, Jin Zhu, Feng Xu, Seunghun Hong, gobardi, Procolo Lucignano, Giacomo Rotoli, Gio- and Chad A. Mirkin, “”dip-pen” nanolithography,” Sci- vanni Piero Pepe, Arturo Tagliacozzo, and Floriana Lom- ence 283, 661–663 (1999). 120 bardi, “Recent achievements on the physics of high-tc su- Elena Bellido, Roc´ıo de Miguel, Daniel Ruiz-Molina, perconductor Josephson junctions: Background, perspec- Anabel Lostao, and Daniel Maspoch, “Controlling the tives and inspiration,” J. Supercond. Nov. Magn. 26, 21– number of proteins with dip-pen nanolithography,” Adv. 41 (2013). Mater. 22, 352–355 (2010). 105 S. A. Cybart, E. Y. Cho, J. T. Wong, B. H. Wehlin, M. K. 121 M. J. Mart´ınez-Pérez,E. Bellido, R. de Miguel, J. Sesé, Ma, C. Huynh, and R. C. Dynes, “Nano Josephson su- A. Lostao, C. Gómez-Moreno, D. Drung, T. Schurig, perconducting tunnel junctions in YBa2Cu3O7−δ directly D. Ruiz-Molina, and F. Luis, “Alternating current mag- patterned with a focused helium ion beam,” Nat. Nan- netic susceptibility of a molecular magnet submonolayer otechnol. 10, 598–602 (2015). directly patterned onto a micro superconducting quantum 106 E. Y. Cho, M. K. Ma, Chuong Huynh, K. Pratt, D. N. interference device,” Appl. Phys. Lett. 99, 032504 (2011). Paulson, V. N. Glyantsev, R. C. Dynes, and Shane A. 122 Elena Bellido, Pablo González-Monje, Ana Repollés, Cybart, “YBa2Cu3O7−δ superconducting quantum inter- Mark Jenkins, Javier Sesé, Dietmar Drung, Thomas ference devices with metallic to insulating barriers written Schurig, Kunio Awaga, Fernando Luis, and Daniel Ruiz- with a focused helium ion beam,” Appl. Phys. Lett. 106, Molina, “Mn12 single molecule magnets deposited on µ- 252601 (2015). SQUID sensors: the role of interphases and structural 107 “Magnetic nanoparticles,” MRS Bulletin 38 (2013), issue modifications,” Nanoscale 5, 12565–12573 (2013). 11. 123 M. Jenkins, D. Gella, A. Repollés,O. Roubeau, G. Arom´ı, 108 J. Bartolomé,F. Luis, and J. F. Fernández,eds., Molec- D. Drung, T. Schurig, M. C. Pallarés,J. Sesé,A. I. Lostao, ular Magnets. Physics and Applications (Springer-Verlag, and F. Luis, (2015), unpublished. Berlin Heidelberg, 2014). 124 A. Buchter, J. Nagel, D. Rüffer,F. Xue, D. P. Weber, 109 Michael N. Leuenberger and Daniel Loss, “Quantum O. F. Kieler, T. Weimann, J. Kohlmann, A. B. Zorin, computing in molecular magnets,” Nature 410, 789–793 E. Russo-Averchi, R. Huber, P. Berberich, A. Fontcuberta (2001). i Morral, M. Kemmler, R. Kleiner, D. Koelle, D. Grundler, 110 Lapo Bogani and Wolfgang Wernsdorfer, “Molecular spin- and M. Poggio, “Reversal mechanism of an individual Ni tronics using single-molecule magnets,” Nature Materials nanotube simultaneously studied by torque and SQUID 7, 179–186 (2008). magnetometry,” Phys. Rev. Lett. 111, 067202 (2013). 111 W. Wernsdorfer, E. Bonet Orozco, K. Hasselbach, 125 A. Buchter, R. Wölbing, M. Wyss, O. F. Kieler, A. Benoit, B. Barbara, N. Demoncy, A. Loiseau, H. Pas- T. Weimann, J. Kohlmann, A. B. Zorin, D. Rüffer, card, and D. Mailly, “Experimental evidence of the Néel- F. Matteini, G. Tütüncüoglu,F. Heimbach, A. Kleib- Brown model of magnetization reversal,” Phys. Rev. Lett. ert, A. Fontcuberta i Morral, D. Grundler, R. Kleiner, 78, 1791–1794 (1997). D. Koelle, and M. Poggio, “Magnetization reversal of an 112 W. Wernsdorfer, K. Hasselbach, A. Benoit, W. Werns- individual exchange-biased permalloy nanotube,” Phys. dorfer, B. Barbara, D. Mailly, J. Tuaillon, J. P. Perez, Rev. B 92, 214432 (2015). V. Dupuis, J. P. Dupin, G. Guiraud, and A. Perex, “High 126 J. R. Kirtley, “Prospects for imaging magnetic nanopar- sensitivity magnetization measurements of nanoscale ticles using a scanning SQUID microscope,” Supercond. cobalt clusters,” J. Appl. Phys. 78, 7192–7195 (1995). Sci. Technol. 22, 064008 (2009). 113 M. Jamet, V. Dupuis, P. Mélinon,G. Guiraud, A. Pérez, 127 Elena Bellido, Neus Domingo abd Isaac Ojea-Jiménez, W. Wernsdorfer, A. Traverse, and B. Baguenard, “Struc- and Daniel Ruiz-Molina, “Structuration and integration ture and magnetism of well defined cobalt nanoparticles of magnetic nanoparticles on surfaces and devices,” Small embedded in a niobium matrix,” Phys. Rev. B 62, 493– 8, 1465–1491 (2012). 499 (2000). 128 Andrea Cornia, Antonio C. Fabretti, Mirko Pacchioni, 114 R. Córdoba,J. Sesé, J.M. De Teresa, and M.R. Ibarra, Laura Zobbi, Daniele Bonacchi, Andrea Caneschi, “High-purity cobalt nanostructures grown by focused- Dante Gatteschi, Roberto Biagi, Umberto Del Pennino, electron-beam-induced deposition at low current,” Micro- Valentina De Renzi, Leonid Gurevich, and Herre S. J. electron. Eng. 87, 1550–1553 (2010). Van der Zant, “Direct observation of single-molecule mag- 115 M J Mart´ınez-Pérez,B Müller,D Schwebius, D Korinski, nets organized on gold surfaces,” Angew. Chem. Int. Ed. R Kleiner, J Sesé, and D Koelle, “Nanosquid magnetom- 42, 1645–1648 (2003). etry of individual cobalt nanoparticles grown by focused 129 Eugenio Coronado, Alicia Forment-Aliaga, Francisco M. electron beam induced deposition,” Supercond. Sci. Tech- Romero, Valdis Corradini, Roberto Biagi, Valentina De nol. 30, 024003 (2017). Renzi, Alessandro Gambardella, and Umberto del Pen- 116 M. Martin, L. Roschier, P. Hakonen, U¨ Parts, M. Paala- nino, “Isolated Mn12 single-molecule magnets grafted on nen, B. Schleicher, and E. I. Kauppinen, “Manipulation gold surfaces via electrostatic interactions,” Inorg. Chem. of Ag nanoparticles utilizing noncontact atomic force mi- 44, 7693–7695 (2005). croscopy,” Appl. Phys. Lett. 73, 1505–1507 (1998). 130 S. K. H. Lam, Wenrong Yang, H. T. R. Wiogo, and 117 Chris I. Pakes, Damien P. George, Sven Ramelow, Al- C. P. Foley, “Attachment of magnetic molecules on a berto Cimmino, David N. Jamieson, and Steven Prawer, nanoSQUID,” Nanotechnology 19, 285303 (2008). 25

131 W. Wernsdorfer, E. Bonet Orozco, K. Hasselbach, 146 M.B. Ketchen, J.R. Kirtley, and M. Bhushan, “Miniature A. Benoit, D. Mailly, O. Kubo, H. Nakano, and B. Bar- vector magnetometer for scanning SQUID microscopy,” bara, “Macroscopic quantum tunneling of magnetization IEEE Trans. Appl. Supercond. 7, 3139–3142 (1997). of single ferrimagnetic nanoparticles of barium ferrite,” 147 Www.hypres.com. Phys. Rev. Lett. 79, 4014–4017 (1997). 148 J. R. Kirtley, C. C. Tsuei, K. A. Moler, V. G. Kogan, J. R. 132 Christophe Thirion, Wolfgang Wernsdorfer, and Do- Clem, and A. J. Turberfield, “Variable sample tempera- minique Mailly, “Switching of magnetization by nonlinear ture scanning superconducting quantum interference de- resonance studied in single nanoparticles,” Nature Mater. vice microscope,” Appl. Phys. Lett. 74, 4011–4013 (1999). 2, 524–527 (2003). 149 J. Mannhart, H. Hilgenkamp, B. Mayer, Ch. Gerber, J. R. 133 D. Le Roy, R. Morel, S. Pouget, A. Brenac, L. Notin, Kirtley, K. A. Moler, and M. Sigrist, “Generation of mag- T. Crozes, and W. Wernsdorfer, “Bistable coupling netic flux by single grain boundaries of YBa2Cu3O7−x,” states measured on single Co nanoclusters deposited on Phys. Rev. Lett. 77, 2782 (1996). CoO(111),” Phys. Rev. Lett. 107, 057204 (2011). 150 C. C. Tsuei, J. R. Kirtley, C. C. Chi., Lock See Yu-Jahnes, 134 W. Wernsdorfer, Magnétométrie à micro-SQUID pour A. Gupta, T. Shaw., J. Z. Sun., and M. B. Ketchen, l’étudede particules ferromagnétiquesisoléesaux échelles “Pairing symmetry and flux quantization in a tricrystal sub-microniques, Ph.D. thesis, Joseph Fourier University, superconducting ring of YBa2Cu3O7−δ,” Phys. Rev. Lett. Grenoble (1996). 73, 593–596 (1994). 135 W. Wernsdorfer and R. Sessoli, “Quantum phase interfer- 151 H. Hilgenkamp, Ariando, H.-J. H. Smilde, D. H. A. Blank, ence and parity effects in magnetic molecular clusters,” G. Rijnders, H. Rogalla, J. R. Kirtley, and C. C. Tsuei, Science 284, 133–135 (1999). “Ordering and manipulation of the magnetic moments in 136 W. Reim, R. H. Koch, A. P. Malozemoff, M. B. Ketchen, large-scale superconducting π-loop arrays,” Nature 422, and H. Maletta, “Magnetic equilibrium noise in spin- 50–53 (2003). 152 glasses: Eu0.4Sr0.6S,” Phys. Rev. Lett. 57, 905–908 Nicholas C. Koshnick, Martin E. Huber, Julie A. Bert, (1986). Clifford W. Hicks, Jeff Large, Hal Edwards, and 137 D. D. Awschalom, J. F. Smyth, G. Grinstein, D. P. Di- Kathryn A. Moler, “A terraced scanning supper con- Vincenzo, and D. Loss, “Macroscopic quantum tunneling ducting quantum interference device susceptometer with in magnetic proteins,” Phys. Rev. Lett. 68, 3092–3095 submicron pickup loops,” Appl. Phys. Lett. 93, 243101 (1992). (2008). 138 David D. Awschalom, David P. DiVincenzo, and 153 P. G. Björnsson,B. W. Gardner, J. R. Kirtley, and K. A. Joseph F. Smyth, “Macroscopic quantum effects in Moler, “Scanning superconducting quantum interference nanometer-scale magnets,” Science 258, 414–421 (1992). device microscope in a dilution refrigerator,” Rev. Sci. 139 F. Luis, A. Repollés,M. J. Mart´ınez-Pérez,D. Aguilà, Instr. 72, 4153–4158 (2001). O. Roubeau, D. Zueco, P. J. Alonso, M. Evangelisti, 154 Katja C. Nowack, Eric M. Spanton, Matthias Baenninger, A. Camón,J. Sesé,L. A. Barrios, and G. Arom´ı,“Molecu- Markus König,John R. Kirtley, Beena Kalisky, C. Ames, lar prototypes for spin-based CNOT and SWAP quantum Philipp Leubner, Christoph Brüne,Hartmut Buhmann, gates,” Phys. Rev. Lett. 107, 117203 (2011). Laurens W. Molenkamp, David Goldhaber-Gordon, and 140 M. J. Mart´ınez-Pérez, S. Cardona-Serra, C. Schlegel, Kathryn A. Moler, “Imaging currents in HgTe quantum F. Moro, P. J. Alonso, H. Prima-Garc´ıa,J. M. Clemente- wells in the quantum spin Hall regime,” Nature Materials Juan, M. Evangelisti, A. Gaita-Ariño,J. Sesé, J. van 12, 787–791 (2013). Slageren, E. Coronado, and F. Luis, “Gd-based single- 155 Julie A. Bert, Beena Kalisky, Christopher Bell, Minu Kim, ion magnets with tunable magnetic anisotropy: Molecu- Yasuyuki Hikita, Harold Y. Hwang, and Kathryn A. lar design of spin qubits,” Phys. Rev. Lett. 108, 247213 Moler, “Direct imaging of the coexistence of ferromag- (2012). netism and superconductivity at the LaAlO3/SrTiO3 in- 141 M. J. Mart´ınez-Pérez, O. Montero, M. Evange- terface,” Nature Physics 7, 767–771 (2011). listi, F. Luis, J. Sesé, S. Cardona-Serra, and 156 Beena Kalisky, Eric M. Spanton, Hilary Noad, John R. E. Coronado, “Fragmenting gadolinium: Mononuclear Kirtley, Katja C. Nowack, Christopher Bell, Hi- polyoxometalate-based magnetic coolers for ultra-low roki K. Sato, Masayuki Hosoda, Yanwu Xie, Yasuyuki temperatures,” Adv. Mater. 24, 4301–4305 (2012). Hikita, Carsten Woltmann, Georg Pfanzelt, Rainer Jany, 142 Christian Cervetti, Angelo Rettori, Maria Gloria Pini, Christoph Richter, Harold Y. Hwang, Jochen Mannhart, Andrea Cornia, Ana Repollés, Fernando Luis, Mar- and Kathryn A. Moler, “Locally enhanced conductivity tin Dressel, Stephan Rauschenbach, Klaus Kern, Marko due to the tetragonal domain structure in LaAlO3/SrTiO3 Burghard, and Lapo Bogani, “The classical and quantum heterointerfaces,” Nature Mater. 12, 1091–1095 (2013). dynamics of molecular spins on graphene,” Nat. Mater. 157 D. J. Hykel, Z. S. Wang, P. Castellazzi, T. Crozes, 15, 164–168 (2016). G. Shaw, K. Schuster, and K. Hasselbach, “MicroSQUID 143 J. R. Kirtley, “SQUID microscopy for fundamental stud- force microscopy in a dilution refrigerator,” J. Low. Temp. ies,” Physica C 368, 55–65 (2002). Phys. 175, 861–867 (2014). 144 J. R. Kirtley, M. B. Ketchen, K. G. Staviasz, J. Z. Sun, 158 C. Veauvy, K. Hasselbach, and D. Mailly, “Micro-SQUID W. J. Gallagher, S. H. Blanton, and S. J. Wind, “High- microscopy of vortices in a perforated superconducting al resolution scanning SQUID microscope,” Appl. Phys. film,” Phys. Rev. B 70, 214513 (2004). Lett. 66, 1138–1140 (1995). 159 C. Paulsen, D. J. Hykel, K. Hasselbach, and D. Aoki, 145 M. Ketchen and J. R. Kirtley, “Design and performance “Observation of the Meissner-Ochsenfeld effect and the aspects of pickup loop structures for miniature SQUID absence of the Meissner state in UCoGe,” Phys. Rev. Lett. magnetometry,” IEEE Trans. Appl. Supercond. 5, 2133– 109, 237001 (2012). 2136 (1995). 26

160 Yonathan Anahory, Jonathan Reiner, Lior Embon, Dorri ductors at nanometer scales,” Scientiﬁc Reports 5, 7598 Halbertal, Anton Yakovenko, Yuri Myasoedov, Michael L. (2015). Rappaport, Martin E. Huber, and Eli Zeldov, “Three- 163 Yonathan Anahory, Lior Embon, Chang Jian Li, Sumilan junction SQUID-on-tip with tunable in-plane and out-of- Banerjee, Alexander Meltzer, Hoovinakatte R. Naren, An- plane magnetic ﬁeld sensitivity,” Nano Letters 14, 6481– ton Yakovenko, Jo Cuppens, Yuri Myasoedov, Michael L. 6487 (2014). Rappaport, Martin E. Huber, Karen Michaeli, Thiru- 161 A. Finkler, D. Vasyukov, Y. Segev, L. Ne’eman, E. O. malai Venkatesan, Ariando, and Eli Zeldov, “Emer- Lachman, M. L. Rappaport, Y. Myasoedov, E. Zeldov, gent nanoscale superparamagnetism at oxide interfaces,” and M. E. Huber, “Scanning superconducting quantum Nat. Commun. (in press) (2016), arXiv:1509.01895 [cond- interference device on a tip for magnetic imaging of mat.str-el], arXiv:1509.01895 [cond-mat.str-el]. nanoscale phenomena,” Rev. Sci. Instr. 83, 073702 (2012), 164 Ella O. Lachman, Andrea F. Young, Anthony Richardella, arXiv:1206.2853v1 [cond-mat.supr-con]. Jo Cuppens, H. R. Naren, Yonathan Anahory, Alexan- 162 L. Embon, Y. Anahory, A. Suhov, D. Halbertal, J. Cup- der Y. Meltzer, Abhinav Kandala, Susan Kempinger, pens, A. Yakovenko, A. Uri, Y. Myasoedov, M. L. Rappa- Yuri Myasoedov, Martin E. Huber, Nitin Samarth, and port, M. E. Huber, A. Gurevich, and E. Zeldov, “Prob- Eli Zeldov, “Visualization of superparamagnetic dynam- ing dynamics and pinning of single vortices in supercon- ics in magnetic topological insulators,” Science Advances 1, e1500740 (2015).