betaFeedback

WorksPhysics and Astronomy

Coherent Electronic Lensing in Proximitized Graphene

Abstract

Spatially resolved control of coherent electron transport offers a route to understanding and engineering interference phenomena in hybrid superconducting devices. Here, we use Scanning Gate Microscopy (SGM) to investigate Fabry-Pérot (FP) interference in the n-p-n cavity of a ballistic graphene Josephson junction. Spatial maps of the superconducting switching current and normal-state conductance reveal circular interference fringes arising from two distinct effects of the SGM tip potential. A global shift of the Fermi wave vector modulates the FP resonances of the cavity, while a local perturbation produces additional oscillations that persist at fixed Fermi wave vector as the tip bias is varied. An analytical model identifies the origin of this local response with coherent electron lensing, in which the tip modifies the phase accumulated along interfering trajectories. The measured fringe patterns in the switching current indicate that this mechanism affects superconducting transport, consistent with tip-induced shifts of the resonance conditions for Andreev bound states. These results establish SGM as a tunable local lens for controlling interference in superconducting ballistic graphene devices.

Electron quantum optics provides a powerful framework for exploring coherent transport in mesoscopic conductors [1, 2], in which the propagation of electrons can be understood in terms of trajectories, refraction, focusing, and interference. The solid-state electronic counterpart of optical phenomena has been introduced in high mobility two-dimensional electron gases (2DEGs), demonstrating coherent electronic beam control, electronic interferometry, and collider-like experiments [3, 4, 5, 6, 7, 8, 9].

In this context, graphene is regarded as a key experimental platform, thanks to the linear dispersion relation and the chiral nature of the charge carriers that give rise to transport phenomena without direct analogues in conventional semiconductor systems [10, 11]. For instance, the possibility of electrostatically defining p-n junctions, which act as interfaces between electron-like and hole-like states, allows for unusual electron-optical functionalities, due to Klein tunneling and negative refraction [12, 13, 14], leading to carrier guiding, focusing, and Veselago lensing [15, 16, 17].

An important manifestation of phase-coherent electron propagation is Fabry-Pérot (FP) interferometry [8, 18, 19], in which carriers undergo multiple reflections between partially transmitting interfaces, producing conductance oscillations as a function of carrier density, gate voltage, and magnetic field. In graphene, such interference can arise in electrostatically defined n-p-n junctions, where the two interfaces form a cavity for charge carriers [20]. There, the interplay between Klein tunneling, which favors transmission at near-normal incidence, and the angle-dependent reflection at the interfaces gives rise to an interference pattern that reflects the chiral nature of graphene quasiparticles [21, 22].

Although conductance measurements reveal the presence of coherent interference, they generally provide only indirect information about the spatial trajectories responsible for it. Scanning Gate Microscopy (SGM) can overcome this limitation. Indeed, by using a biased, movable tip to locally perturb the electrostatic potential of a device, local maps of conductance can be recorded. In conventional 2DEGs, SGM has proven useful to image coherent electron flow, to enable local carrier manipulation, and to build interferometric setups [23, 24, 25, 26, 27, 28, 29]. In graphene, previous work has used SGM to investigate coherent transport, demonstrating guided electronic states, disorder-induced phenomena, and electron-optical phenomena [30, 31, 32, 33, 34, 35, 36, 37, 38, 39].

Introducing superconducting contacts adds a further dimension to this picture. Graphene Josephson junctions support phase-coherent transport, in which supercurrent is mediated by Andreev processes: an electron incident on a superconducting interface is reflected as a hole, while a Cooper pair enters the superconductor. The resulting electron–hole dynamics provides a route to studying superconducting proximity effects in a material whose carrier density and polarity remain electrostatically tunable. Experiments on graphene Josephson junctions have demonstrated ballistic supercurrent transport, quantum oscillations of the critical current, and FP interference in the superconducting state [40, 41, 42]. FP oscillations have been shown to modulate the supercurrent, with local maxima (minima) in the switching current aligned with local maxima (minima) in the conductance [40, 41, 42, 43]. However, how the spatial structure of coherent transport evolves when normal-state electron optics is coupled to superconducting correlations is still an open question.

In this context, spatially resolved measurements are particularly valuable because they can provide information that is not readily accessible from global transport characteristics alone [44, 45, 46]. Recently, SGM has been successfully applied to imaging defects in superconducting resonators [47] and Andreev reflections in graphene [48], as well as modulating supercurrent flow in Josephson junctions [49, 50].

Here, we manipulate Fabry-Pérot interference in a ballistic graphene Josephson junction employing SGM. Differently from previous SGM experiments on graphene-based devices, where the tip was used to create an electrostatic cavity [36, 38], in our work the SGM tip is used to actively control the properties of the n-p-n cavity that already exists in the sample. We experimentally observe a series of spatial circular fringes in the switching current, correlating with analogous patterns in the normal-state conductance. By combining bias and gate-dependent measurements with analytical modeling, we reveal that the fringes are generated by two effects. First, a global tip-induced renormalization of the Fermi wave vector shifts the FP oscillations. Second, the local nature of the tip potential produces oscillations at fixed Fermi wave vector, arising from a coherent lensing effect. Interestingly, the same mechanism accounts also for the fringe pattern observed in the superconducting regime, where Andreev bound states and the Fabry-Pérot cavity coexist. The tip locally modifies the quantum phase of interfering trajectories and thus allows to control quantum-coherent transport across hundreds of nanometers both in the normal and superconducting state.

We investigate a ballistic graphene Josephson junction fabricated with Niobium (Nb) superconducting edge-contacts. We employ high quality graphene crystals grown by Chemical Vapor Deposition (CVD) [51, 52] and encapsulated in 30 nm-thick hexagonal Boron Nitride (hBN) flakes. The hBN-graphene-hBN heterostructure is placed on a Si/SiO2_2 substrate that acts as a back-gate via the application of a voltage VBGV_{\mathrm{BG}}. Fabrication of the device follows the procedure described in Refs. [43, 50, 53] and briefly summarized in the Supporting Information. A three-dimensional sketch of the Josephson junction and of the measurement setup is reported in Figure 1a. An optical microscopy image of the device is shown in Figure 1b.

Figure 1. (a) Three-dimensional schematics of the graphene Josephson junction and SGM setup. Graphene is encapsulated in two hBN flakes and edge-contacted by Nb superconducting leads. The biased SGM tip scans above the device at fixed distance. (b) Optical microscopy image of the investigated device (junction length L=600L = 600 nm and width W=3W = 3 µm). The black rectangle indicates the scanning area of the maps shown in panels c and e. (c) SGM map of switching current IswI_{\mathrm{sw}} as a function of the SGM tip coordinates. Tip bias Vtip=10V_{\mathrm{tip}} = 10 V, back-gate voltage VBG=−20V_{\mathrm{BG}} = -20 V. The white dot and diamond indicate the position of the VV-IbiasI_{\mathrm{bias}} curves shown in panel d. (d) Representative VV-IbiasI_{\mathrm{bias}} curves from map in panel c at ytip=0y_{\mathrm{tip}} = 0 and different xtipx_{\mathrm{tip}} coordinates, showing the tip-induced modulation of IswI_{\mathrm{sw}} and GG. (Upper panel) xtip=−1.54x_{\mathrm{tip}} = -1.54 µm, white dot in Figure 1c. (Lower panel) xtip=−1.21x_{\mathrm{tip}} = -1.21 µm, white diamond in Figure 1c. IswI_{\mathrm{sw}} (black vertical dashed lines) is obtained by setting a threshold on the measured voltage drop VV, while GG is obtained from a best-fit in the dissipative branch of the VV-IbiasI_{\mathrm{bias}} curves (red dashed lines). (e) SGM map of junction conductance GG obtained from the same dataset of panel c, by fitting the dissipative branch of VV-IbiasI_{\mathrm{bias}} curves. (f) Representative linecuts of IswI_{\mathrm{sw}} (upper panel) and of GG (lower panel) from panels c and e, obtained as the average of three lines around ytip=0y_{\mathrm{tip}} = 0, showing that IswI_{\mathrm{sw}} and GG share the same oscillatory pattern. Dotted vertical lines highlight the alignment between local maxima and minima in IswI_{\mathrm{sw}} and GG.

All measurements presented in this work were performed in a Janis cryostat equipped with an Attocube AFM head operating at T=300T = 300 mK. Electrical lines for sample measurement and to AFM components are equipped with cryogenic filters to preserve the superconducting properties of the investigated device. The SGM maps reported here are obtained by working in flying mode: the tip is scanned above the sample in non-contact mode at a constant tip-to-graphene distance of d=100d = 100 nm (corresponding to a tip-to-sample distance d=70d = 70 nm, considering the thickness of 30 nm of the top hBN flake). By applying a voltage bias VtipV_{\mathrm{tip}}, the tip acts as a movable local gate.

In Figure 1c we report the switching current IswI_{\mathrm{sw}} of the device as a function of tip position (xtip,ytip)(x_{\mathrm{tip}}, y_{\mathrm{tip}}), acquired by setting the back-gate voltage to VBG=−20V_{\mathrm{BG}} = -20 V (corresponding to a hole carrier density n=8.2×1011n = 8.2 \times10^{11} cm−2^{-2}), and applying a positive tip voltage bias Vtip=10V_{\mathrm{tip}} = 10 V. The map was obtained by measuring VV-IbiasI_{\mathrm{bias}} curves at each pixel of a grid defined in the 5×3.65 \times3.6 μm2^2 area indicated by the black rectangle in

Figure 1b. Each VV-IbiasI_{\mathrm{bias}} curve was obtained by sweeping the current bias IbiasI_{\mathrm{bias}} from 0 to 400 nA, and the switching current IswI_{\mathrm{sw}} was extracted by setting a threshold on the measured voltage drop VV.

While scanning over the junction, the biased SGM tip modulates IswI_{\mathrm{sw}}. Thanks to the high resolution of the SGM map of Figure 1c, two contributions in the tip-induced modulation of IswI_{\mathrm{sw}} can be distinguished. First, a reduction of IswI_{\mathrm{sw}} is observed, which is maximum when the tip is moved to the junction center. Similar behavior has been recently reported in Refs. [49, 50]. The tip creates a potential modulation that is analogous to that of an additional top gate with a different (smaller, in the experimental conditions employed here) lever arm compared to the back-gate. In the hole doping regime and for positive VtipV_{\mathrm{tip}}, a reduction of the supercurrent is observed in Figure 1c. In the following we refer to this effect as the background modulation of IswI_{\mathrm{sw}}.

Interestingly, in addition to the background modulation described above, we also observe multiple circular fringes that intersect each other. This additional tip-induced effect on IswI_{\mathrm{sw}} can be clearly appreciated in the representative VV-IbiasI_{\mathrm{bias}} curves shown in Figure 1d. These curves are taken at ytip=0y_{\mathrm{tip}} = 0, at different xtipx_{\mathrm{tip}} coordinates: on a fringe maximum (xtip=−1.54x_{\mathrm{tip}} = -1.54 µm, white dot in Figure 1c), giving Isw=225I_{\mathrm{sw}} = 225 nA (top panel), and on a fringe minimum (xtip=−1.21 μmx_{\mathrm{tip}}=-1.21\ \mu\mathrm{m}, white diamond in Figure 1c), yielding Isw=193 nAI_{\mathrm{sw}}=193\ \mathrm{nA} (bottom panel). A difference of ΔIsw=32 nA\Delta I_{\mathrm{sw}}=32\ \mathrm{nA} is observed between the two curves. An additional IswI_{\mathrm{sw}} map obtained with opposite tip bias as in Figure 1c shows comparable features and is presented in the Supporting Information.

To further investigate these spatial features, it is useful to relate maps of IswI_{\mathrm{sw}} to the corresponding ones in the dissipative regime, i.e., at Ibias>IswI_{\mathrm{bias}}>I_{\mathrm{sw}}. Indeed, from the slope of measured VV-IbiasI_{\mathrm{bias}} curves it is also possible to extract the sample resistance RR, obtaining the conductance as G=1/RG=1/R. In Figure 1d, linear fits in the current interval 300−400 nA300-400\ \mathrm{nA} are shown as red dashed lines. A larger G=110.2±0.1 e2/hG=110.2\pm0.1\ e^2/h is obtained from the VV-IbiasI_{\mathrm{bias}} curve corresponding to the larger IswI_{\mathrm{sw}} (top panel), as compared to G=106.5±0.1 e2/hG=106.5\pm0.1\ e^2/h obtained from the second VV-IbiasI_{\mathrm{bias}} curve (bottom panel). By repeating the same process for all curves, we obtain the SGM map of the conductance GG shown in Figure 1e. In this map we observe a clear spatial modulation analogous to that pertaining to IswI_{\mathrm{sw}}. Multiple circular intersecting fringes are present, which share the same spatial pattern as in the IswI_{\mathrm{sw}} map. A global background modulation is also observed, with minimum conductance with the tip above the junction center.

It is well known that in Josephson junctions a positive correlation exists between normal-state resistance RnR_{\mathrm{n}} and switching current IswI_{\mathrm{sw}}. While both quantities are gate-tunable, the Isw⋅RnI_{\mathrm{sw}}\cdot R_{\mathrm{n}} product is observed to be constant and independent of back-gate voltage [40, 41]. This direct link is also observed here in the switching current reported in Figure 1c and in the conductance map of Figure 1e. Here we take the measured R=1/GR=1/G as an approximation for RnR_{\mathrm{n}}, noting that this is an underestimation which neglects the excess current (indeed, in the chosen parameter range, VV remains much smaller than the gap of the superconductor, resulting in an enhancement of GG with respect to the normal state [54]). This direct link is further clarified in Figure 1f, where we compare representative line-cuts of IswI_{\mathrm{sw}} and GG. Local maxima and minima in these two plots are aligned, as exemplified by the guiding black dashed lines. For this reason, it is equivalent to consider either the sample conductance or the switching current when investigating the origin of the fringes observed in Figure 1c,Figure 1e. In the following we focus on the normal-state conductance GG, measured with a lock-in amplifier applying a 100 nA100\ \mathrm{nA} AC bias on top of a fixed DC bias of 2 μA2\ \mu\mathrm{A} and measuring the voltage drop. Compared to VV-IbiasI_{\mathrm{bias}} curves, this technique allows not only for much faster measurements, but these are also characterized by a lower level of noise. In the Supporting Information we demonstrate that measurements performed either by extracting the sample conductance from VV-IbiasI_{\mathrm{bias}} curves (as shown in Figure 1e) or with a lock-in amplifier give consistent results and share the same oscillatory pattern.

To understand the nature of the observed fringe pattern it is instructive to consider first the conductance behavior in absence of the tip. In Figure 2a we show a back-gate sweep (junction conductance GG as a function of back-gate voltage VBGV_{\mathrm{BG}}) measured without the tip. As reported for previous examples of edge-contacted graphene Josephson junctions [40, 41, 43, 55], this curve is clearly asymmetric, with a lower conductance on the left side of the charge neutrality point CNP (hole doping) compared to the right side (electron doping). This asymmetry originates from the interplay between Fermi level pinning at graphene-Nb interfaces and doping induced by the back-gate. The graphene region in proximity to the Nb contacts is always electron-doped regardless of the applied VBGV_{\mathrm{BG}} due to Fermi level pinning, while the doping in the central region of the junction is controlled by the back-gate. As a result, an n-p-n cavity forms in the device for p-type doping in the central part of the junction. This electrostatic cavity behaves as a FP interferometer, with its interfaces acting as partially reflecting mirrors for the massless Dirac carriers, leading to interference among the internally reflected trajectories [21] (as sketched in the lower inset of Figure 2a). By varying the depth of the cavity, its resonance condition changes, leading to an alternation between constructive and destructive interference, which results in conductance oscillations versus back-gate. The upper inset of Figure 2a highlights the presence of this quantum coherent phenomenon in our device.

Figure 2. (a) Back-gate sweep (conductance GG as a function of back-gate voltage VBGV_{\mathrm{BG}}). The charge neutrality point (CNP) is shifted to negative back-gate voltages (VBGCNP=−6.3V_{\mathrm{BG}}^{\mathrm{CNP}}=-6.3 V in this specific device). (Upper inset) FP oscillations: zoom-in of junction conductance variation ΔG\Delta G as a function of VBGV_{\mathrm{BG}} in the range (−20,−13-20,-13) V, obtained by subtracting a polynomial background from the conductance GG of the main panel. (Lower inset) Pictorial sketch of FP interference in the n-p-n cavity. LcL_{\mathrm{c}} is the length of the p-doped region. (b) Map of conductance GG as a function of back-gate voltage VBGV_{\mathrm{BG}} and tip bias VtipV_{\mathrm{tip}}. Tip is positioned at junction center, as indicated by the black star in the inset. The black dashed line represents the fitted position of the effective charge neutrality point. The black parallelogram corresponds to the data shown in panel d. The conductance GG is measured with a lock-in amplifier by applying a bias large enough to suppress any contribution from the excess current. (c) FP oscillations in the hole doping regime as a function of VBGV_{\mathrm{BG}} and VtipV_{\mathrm{tip}}: conductance variation ΔG\Delta G obtained from panel b by subtracting a two-dimensional polynomial background. (d) Conductance variation ΔG\Delta G as a function of effective Fermi wave vector kF,effk_{\mathrm{F,eff}} (see main text) and VtipV_{\mathrm{tip}}. The displayed area corresponds to the black parallelogram in panel b. (e) Line-cuts of ΔG\Delta G as a function of kF,effk_{\mathrm{F,eff}} and (f) corresponding Fourier transform for Vtip=10V_{\mathrm{tip}}=10 V (black), Vtip=0V_{\mathrm{tip}}=0 V (blue), Vtip=−10V_{\mathrm{tip}}=-10 V (grey). Position of line-cuts is also indicated by horizontal colored arrows in panel d. (g) Comparison between the analytical model of Eq. (2) and experimental line-cuts as a function of VtipV_{\mathrm{tip}} extracted at fixed kF,effk_{\mathrm{F,eff}} from panel d (as indicated by vertical colored arrows in panel d). The same normalization has been applied to all curves, and a background has been subtracted from GG to isolate ΔG\Delta G, as in panels c,d. The model parameters used are η=0.1\eta=0.1 V−1^{-1}, Lc=530L_{\mathrm{c}}=530 nm, W=3W=3 μ\mum, R=430R=430 nm, ytip=0y_{\mathrm{tip}}=0.

We now discuss the dependence of the conductance on VtipV_{\mathrm{tip}}, starting by keeping the tip position fixed. In Figure 2b we report the sample conductance GG as a function of back-gate voltage VBGV_{\mathrm{BG}} and tip voltage VtipV_{\mathrm{tip}}, obtained by sweeping VBGV_{\mathrm{BG}} from −20 V-20\ \mathrm{V} to 0 V0\ \mathrm{V} at different values of VtipV_{\mathrm{tip}} ranging from +10 V+10\ \mathrm{V} to −10 V-10\ \mathrm{V}. The tip position is kept fixed above the junction center at a vertical distance d=100 nmd=100\ \mathrm{nm} from the graphene channel (as indicated in the inset to Figure 2b). In analogy with the CNP in the back-gate sweep in Figure 2a, the conductance minimum in each line at finite VtipV_{\mathrm{tip}} can be identified as an effective CNP. The effective CNP shifts as a function of VtipV_{\mathrm{tip}}, with a linear dependence indicated by the black dashed line in Figure 2b (obtained by fitting points of minimum conductance in each line at fixed VtipV_{\mathrm{tip}}). This linear shift can be attributed to a global effective shift of the chemical potential, in which the tip, when positioned above the junction center, acts as a top-gate with a lever arm of ∼29%\sim29\% relative to the back-gate.

In addition, a series of inclined oscillating features are visible to the left of the effective CNP. This phenomenology can be better seen in Figure 2c, where we report the detrended conductance ΔG\Delta G in the hole doping regime for −20-20 V ≤VBG≤−13.5\le V_{\mathrm{BG}} \le-13.5 V, obtained from the data in Figure 2b by removing a polynomial background (see details in the Supporting Information). Clearly, the sample conductance oscillates as both back-gate voltage and tip bias are changed. The line at Vtip=0V_{\mathrm{tip}} = 0 V nominally corresponds to the back-gate sweep shown in Figure 2a (top inset): comparable amplitude and oscillation period are observed. By applying a finite VtipV_{\mathrm{tip}}, the oscillating pattern does not change abruptly, but smoothly evolves. By inspecting Figure 2c it becomes clear that the tip action cannot be simply reduced to that of a global effective top-gate. Indeed, varying VBGV_{\mathrm{BG}} and VtipV_{\mathrm{tip}} along a diagonal parallel to the effective CNP line leaves the effective chemical potential unchanged. Hence, if the tip could be mapped into a global effective top-gate, features in Figure 2c would appear as a series of inclined but continuous stripes parallel to the effective CNP line, without any crossings (see for example Refs. [21, 22] where multiple gates are used to investigate FP oscillations). This is not the case here, as maxima and minima exhibit multiple crossings along diagonal lines of constant effective chemical potential, with maxima evolving into minima and vice-versa. This additional behavior suggests that the local nature of the tip-induced perturbation plays a crucial role and has to be taken into proper consideration.

As a first hypothesis, one could think that the described pattern originates from a modification of the cavity length LcL_{\mathrm{c}} induced by the tip potential. To test this hypothesis, we extracted the cavity length as a function of VtipV_{\mathrm{tip}}. To account for the global contribution of the tip discussed above, in Figure 2d we plot the ΔG\Delta G data as a function of kF,effk_{\mathrm{F,eff}}. The effective Fermi wave vector kF,effk_{\mathrm{F,eff}} is calculated relative to the position of the effective CNP (black dashed line in Figure 2b), kF,eff=πα(VBG−VCNP,eff)k_{\mathrm{F,eff}} = \sqrt{\pi\alpha(V_{\mathrm{BG}} - V_{\mathrm{CNP,eff}})}, where α\alpha is the back-gate lever arm. Data in Figure 2d correspond to a parallelogram in Figure 2b with lateral edges parallel to the effective CNP position (indicated in Figure 2b by black solid lines). Representative lines of ΔG\Delta G vs kF,effk_{\mathrm{F,eff}} are shown in Figure 2e for different values of VtipV_{\mathrm{tip}} (Vtip=10,0,−10V_{\mathrm{tip}} = 10, 0, -10 V respectively in the top, middle, and bottom panels). An oscillating behavior similar to that observed on the hole doping side of the back-gate sweep (inset to Figure 2a) is observed, showing no relevant variations in terms of oscillation amplitude or frequency for different values of VtipV_{\mathrm{tip}}. We extract the cavity length for all values of VtipV_{\mathrm{tip}} by means of Fourier transform (representative FFT spectra corresponding to line-cuts in Figure 2e are reported in Figure 2f) and obtain Lc=460±25L_{\mathrm{c}} = 460 \pm25 nm. This value shows little to no variation as VtipV_{\mathrm{tip}} is changed over a wide range of 20 V, and is compatible with the cavity length value extracted from the back-gate sweep of Figure 2a, Lc=427±30L_{\mathrm{c}} = 427 \pm30 nm. This observation rules out the renormalization of LcL_{\mathrm{c}} as an explanation for the tip-induced modulation of the conductance at fixed kF,effk_{\mathrm{F,eff}}. The only remaining interpretation is an action of the tip on the coherent FP interference of ballistic trajectories within the cavity.

To understand the interplay between the tip and the cavity, we developed an analytical approach that generalizes, by taking the tip into account, the one that demonstrates the presence of FP oscillations in a graphene p-n-p junction employed by Shytov et al. in Ref. [20]. In the absence of the tip, the two-dimensional scattering problem reduces to a one-dimensional one, since, assuming translational invariance along the transverse coordinate xx the scattering states can be labeled by a well-defined transverse momentum component kxk_x. To calculate the transmission probability in the effective 1D system, Shytov et al. sum the amplitudes of all possible FP interference paths, obtaining a kxk_x-dependent transmission of the form

T(kx)=(1+F(kx)sin⁡2[θ(kx)])−1.T(k_x) = \left(1 + F(k_x)\sin^2[\theta(k_x)]\right)^{-1}.

Here F=4r/(1−r)2F = 4r/(1-r)^2 is a finesse coefficient, depending on the probability rr of internal reflection at the interfaces. The angle θ\theta is the semiclassical phase accumulated along a path between the two interfaces. This approach can be generalized to account for the presence of the tip, which is generally described by a Lorentzian shape (details are provided in the Supporting Information). The potential generated by the tip at a generic point of coordinates (x,y)(x,y), with tip located in (xtip,ytip)(x_{\mathrm{tip}},y_{\mathrm{tip}}), can be written as

Vtip(x,y)≃⟨Vtip⟩+vtip(x,y),(1)V_{\mathrm{tip}}(x,y) \simeq\langle V_{\mathrm{tip}}\rangle+ v_{\mathrm{tip}}(x,y), \tag*{(1)}

where ⟨…⟩\langle\ldots\rangle denotes the spatial average of the Lorentzian profile over the junction, and vtipv_{\mathrm{tip}} represents a local contribution of Lorentzian shape, but of much smaller amplitude than the nominal value of VtipV_{\mathrm{tip}}. The term ⟨Vtip⟩\langle V_{\mathrm{tip}}\rangle is responsible for a global renormalization of the Fermi momentum, which becomes the effective quantity kF,effk_{\mathrm{F,eff}} and manifests as a shift of the effective CNP, as shown in Figure 2. The effect of the Lorentzian perturbation vtip(x,y)v_{\mathrm{tip}}(x,y), by contrast, is local: it deforms the semiclassical phase θ\theta into θ+δθ(x−xtip)\theta+ \delta\theta(x - x_{\mathrm{tip}}), where δθ\delta\theta depends on the impact parameter x−xtipx - x_{\mathrm{tip}} of the electrons relative to the tip center. This is, in all respects, a coherent lensing effect [56], in which the tip actively modifies the quantum phase accumulated along each FP path between the two interfaces.

Within the semiclassical approximation [20] for the transmission probability at the interfaces, one gets r=1−exp⁡(−2kx2/kF,eff2)r = 1 - \exp(-2k_x^2/k_{\mathrm{F,eff}}^2), resulting in an effective finesse Feff(kx,kF,eff)F_{\mathrm{eff}}(k_x,k_{\mathrm{F,eff}}). The expression for the conductance thus becomes (see Supporting Information for more details)

G=4e2h∫dkx2π∫−W/2W/2dx11+Feff(kx,kF,eff)sin⁡2[kF,effLc+δθ(x−xtip)](2)G = \frac{4e^2}{h}\int\frac{d k_x}{2\pi}\int_{-W/2}^{W/2} dx \frac{1}{1+F_{\mathrm{eff}}(k_x,k_{\mathrm{F,eff}})\sin^2[k_{\mathrm{F,eff}}L_c+\delta\theta(x-x_{\mathrm{tip}})]} \tag*{(2)}

where the explicit expression for δθ\delta\theta reads

δθ(x−xtip)=2(ηVtip)kF,effR2R2+(x−xtip)2arctan⁡(L2R2+(x−xtip)2)(3)\delta\theta(x-x_{\mathrm{tip}})=2(\eta V_{\mathrm{tip}})k_{\mathrm{F,eff}}\frac{R^2}{\sqrt{R^2+(x-x_{\mathrm{tip}})^2}}\arctan\left(\frac{L}{2\sqrt{R^2+(x-x_{\mathrm{tip}})^2}}\right) \tag*{(3)}

with ηVtip\eta V_{\mathrm{tip}} the dimensionless fraction of Fermi energy induced by the perturbation vtipv_{\mathrm{tip}} and RR the Lorentzian radius of the tip potential. This coherent phase shift gives rise to a complex oscillation pattern that closely reproduces what is experimentally observed in conductance line cuts as a function of VtipV_{\mathrm{tip}} at fixed kF,effk_{\mathrm{F,eff}}. This is shown in Figure 2g for three different values of kF,effk_{\mathrm{F,eff}}, indicated by the vertical colored arrows in Figure 2d. An excellent agreement is obtained between the experimental curves (dashed lines) and the theoretical ones (solid lines, considering the tip located at junction center). The main harmonic of the oscillations arises from the single-mode contribution at fixed xx, whose transmission amplitude follows a form of the type [1+Fsin⁡2(a+bVtip)]−1[1+F\sin^2(a+bV_{\mathrm{tip}})]^{-1}. We note that finer oscillations are also superimposed on the main harmonic component. These additional oscillations and the deformation of the simple single-mode curve are due to the intricate integral superposition of all transmissions for trajectories originating from different xx values. Moreover, analyzing the dependence of the conductance oscillations on the tip radius, the main harmonic increases systematically as the Lorentzian radius decreases, which supports the lensing interpretation of the incident electron beam (see Supporting Information).

This analytical model not only is able to explain the nature of the observed oscillations, but also demonstrates that the SGM tip departs from the role of a passive probe, effectively becoming an active agent for the coherent control of the quantum phase of electrons in graphene.

To further explore the tip lensing effect, and eventually connect the observed modulation of FP interference under the combined effect of VBGV_{\mathrm{BG}} and VtipV_{\mathrm{tip}} (shown in Figure 2) to the spatial maps analyzed in Figure 1, we next study the conductance as a function of VtipV_{\mathrm{tip}} and xtipx_{\mathrm{tip}}. This intermediate step fixes one of the voltage parameters of Figure 2 (VBG=−20V_{\mathrm{BG}}=-20 V) and one of the spatial coordinates of Figure 1 (ytip=0y_{\mathrm{tip}}=0). The resulting map is shown in Figure 3a. Also in this case we observe a global background reduction (enhancement) of GG with positive (negative) VtipV_{\mathrm{tip}}, arising from the effective shift of the CNP induced by the ⟨Vtip⟩\langle V_{\mathrm{tip}}\rangle term in Eq. (1). Along xtipx_{\mathrm{tip}}, the background modulation becomes less pronounced as the tip moves toward the edges of the junction and ⟨Vtip⟩\langle V_{\mathrm{tip}}\rangle decays. On top of the background, an intricate pattern evolves as a function of both xtipx_{\mathrm{tip}} and VtipV_{\mathrm{tip}}, as a result of tip-induced manipulation of FP interference. Indeed, varying VtipV_{\mathrm{tip}} not only modifies kF,effk_{\mathrm{F,eff}}, but also changes the local term vtipv_{\mathrm{tip}} that enters the lensing phase δθ\delta\theta and ultimately determines the measured conductance (see Eqs. (2),(3)). Similarly, displacing the tip along xtipx_{\mathrm{tip}} simultaneously modifies kF,effk_{\mathrm{F,eff}} and δθ\delta\theta. The interplay between these two effects gives rise to the pattern shown in Figure 3a, and demonstrates the broad parameter space available for coherent control of electron motion using the SGM technique. In Figure 3b, we compare the line-cut at fixed Vtip=10V_{\mathrm{tip}}=10 V extracted from Figure 3a (green horizontal arrow), with the corresponding line at ytip=0y_{\mathrm{tip}}=0 extracted from the SGM map shown in Figure 3f (blue horizontal arrow). These curves share the same oscillatory behavior, demonstrating that the fringe pattern observed in SGM maps is ascribable to tip-tuned FP interference.

SGM maps and simulations of junction conductance

Figure 3. (a) Map of junction conductance GG as a function of VtipV_{\mathrm{tip}} and xtipx_{\mathrm{tip}}. Tip is scanned in the xtipx_{\mathrm{tip}}-direction along a line at ytip=0y_{\mathrm{tip}} = 0, as indicated in the inset. VBG=−20V_{\mathrm{BG}} = -20 V. (b) Comparison of line cuts of GG vs. xtipx_{\mathrm{tip}}. Green line is a horizontal line cut extracted from panel a at fixed Vtip=10V_{\mathrm{tip}} = 10 V (green arrow), blue line is a horizontal line cut extracted from the SGM map of Figure 3f at fixed ytip=0y_{\mathrm{tip}} = 0 (blue arrow). The dashed guiding lines highlight that the curves share the same oscillatory pattern. (c-f) SGM maps of the junction conductance GG in the hole doping regime (VBG=−20V_{\mathrm{BG}} = -20 V), for different values of VtipV_{\mathrm{tip}}. (g-h) Zoom-in of the area in black rectangles in panels e-f. Dashed black lines outline the junction geometry. (i) Simulation of the SGM data in g, for Vtip=7.5V_{\mathrm{tip}} = 7.5 V. (j) Simulation of the SGM data in h, for Vtip=10V_{\mathrm{tip}} = 10 V. The same color scale is used for experimental data (panels a and c-h) and simulations (panels i and j).

Next, we investigate the evolution of the interference fringes at different VtipV_{\mathrm{tip}} values. SGM maps of GG for VtipV_{\mathrm{tip}} ranging from −10-10 V to +10+10 V are shown in Figure 3c to Figure 3f. In all maps, fringe interference patterns are observed, superimposed on a background enhancement or reduction of GG, depending on the sign of VtipV_{\mathrm{tip}}. The fringe patterns are qualitatively similar, but the specific spatial evolution of the fringes varies with VtipV_{\mathrm{tip}}. The formation of circular fringes can be explained by considering the effect of the spatially averaged term ⟨Vtip⟩\langle V_{\mathrm{tip}}\rangle, which is a function of the tip position, ⟨Vtip⟩=⟨Vtip(xtip,ytip)⟩\langle V_{\mathrm{tip}}\rangle=\langle V_{\mathrm{tip}}(x_{\mathrm{tip}},y_{\mathrm{tip}})\rangle. Since these coordinates correspond to the pixel coordinates of the experimental maps, each pixel is associated with a different renormalization of the Fermi energy. Equipotential lines of the ⟨Vtip(xtip,ytip)⟩\langle V_{\mathrm{tip}}(x_{\mathrm{tip}},y_{\mathrm{tip}})\rangle term have a circular shape, which means that along such lines the renormalization of the Fermi momentum is the same. This gives rise to the circular shape of conductance fringes observed experimentally in Figure 3c to Figure 3f and Figure 1c,Figure 1e. However, the conductance is crucially altered by the local lensing effect, which modifies the interference conditions and therefore makes each specific pattern measured by SGM deviate from the equipotential curves of ⟨Vtip⟩\langle V_{\mathrm{tip}}\rangle. To further verify the lensing effect introduced by the tip, we performed numerical simulations of the SGM maps in Figure 3g (Vtip=7.5V_{\mathrm{tip}} = 7.5 V) and Figure 3h (Vtip=10V_{\mathrm{tip}} = 10 V), shown respectively in Figure 3i and Figure 3j. These simulations are based on a coarse grained model on a discretized lattice [49, 50, 57], with the action of the tip modeled by an effective Lorentzian-shaped potential Vtip(r)=Vtip/(1+∣r−rtip∣2/R2)V_{\mathrm{tip}}(\mathbf{r}) = V_{\mathrm{tip}}/(1 + \lvert\mathbf{r} - \mathbf{r}_{\mathrm{tip}} \rvert^{2}/R^{2}) [58]. Details are provided in the Supporting Information. Besides the global background modulation of GG, the numerical simulations effectively reproduce the fringe pattern observed in the experimental maps. This allows us to exclude other non-tip-induced effects as the origin of our experimental observations.

Strikingly, the lensing mechanism determines also the pattern observed in Figure 1c for the switching current when the Josephson junction operates in the superconducting regime. In fact, when the junction supports a supercurrent, Andreev bound states form between the Niobium leads. By resonating at the specific superconducting phase difference across the junction, they enable dissipation-less charge transfer. As discussed in [41], if the electron and hole trajectories cannot form closed loops, phase coherence is lost, and the corresponding Andreev bound state is destroyed, reducing the maximum sustainable supercurrent, i.e., the switching current. The coherent lensing can precisely reproduce this effect. Namely, the tip modifies the phase of both the traveling electrons and holes, and if the superconducting phase difference cannot compensate sufficiently, the tip drives the corresponding Andreev bound state off resonance. This tip-induced off-resonance condition is precisely the same as in FP trajectory lensing, the only difference being that in the normal state a single electron bounces within the cavity. This correspondence shows that Scanning Gate Microscopy can actively control the quantum phase in a wide range of mesoscopic phenomena, spanning from coherent electronic transport to superconductivity.

Fabrication methods

The fabrication of graphene Josephson junctions follows the procedure described in Refs. [43, 50, 53]. We employ the dry pick-up method [59, 60] to assemble the hBN-graphene-hBN heterostructure. The heterostructure comprises two 30 nm-thick hBN flakes and is released on a Si/SiO2_2 substrate that acts as a back-gate. We employ graphene crystals grown by chemical vapor deposition (CVD) [51, 61]. Prior to patterning, the heterostructure is analyzed via Raman spectroscopy and atomic force microscopy (AFM), to identify flat areas with minimal nanoscale strain variations (evaluated from the width of the 2D Raman peak) [52] for device patterning.

A first electron-beam lithography (EBL) patterning step defines a mask used for etching of the heterostructure and deposition of Niobium (Nb) contacts through magnetron sputtering. This step leads to self-aligned one-dimensional contacts [41], whose height matches the thickness of the stack, 60 nm. A second EBL step is used to define the device mesa through a second etching step.

Navigation markers comprising arrows and alphanumeric markers are co-patterned during the first EBL step. As a result, they have the same height as the Nb contacts and the heterostructure, preventing interference during SGM scans. Markers are necessary to navigate on the sample when the system is at T=300T = 300 mK, because no imaging tools are available, other than topography maps. The markers pitch is optimized in such a way that at least one marker will always be present in the 8×88 \times8 μ\mum2^2 scanning area at T=300T = 300 mK.

Additional SGM maps of IswI_{\mathrm{sw}}

In Figure S1 we present an additional map of the switching current IswI_{\mathrm{sw}} as a function of the tip position (xtip,ytip)(x_{\mathrm{tip}}, y_{\mathrm{tip}}), at the same back-gate voltage value as in the main text, VBG=−20V_{\mathrm{BG}} = -20 V, but opposite tip bias, Vtip=−10V_{\mathrm{tip}} = -10 V. As expected for a negative tip bias, we observe an enhancement of the switching current as the tip is moved toward the junction center. Furthermore, similarly to Figure 1c of the main text, circular fringes of approximately constant switching current are present, superimposed to the global background modulation.

Figure S1. SGM map of switching current IswI_{\mathrm{sw}} as a function of the SGM tip coordinates. Tip bias Vtip=−10V_{\mathrm{tip}} = -10 V, back-gate voltage VBG=−20V_{\mathrm{BG}} = -20 V, ΔVBG=−13.7\Delta V_{\mathrm{BG}} = -13.7 V (p-type doping regime).

Correlation of lock-in maps with DC maps

In the main text we show that the switching current IswI_{\mathrm{sw}} and the conductance GG obtained from VV-IbiasI_{\mathrm{bias}} curves share the same spatial modulation and the same fringe pattern, as further exemplified by the linecut comparison in Figure 1f.

Here we demonstrate that measurements of the junction conductance GG performed either by extracting this value from DC bias measurements (i.e., from VV-IbiasI_{\mathrm{bias}} curves) or by measuring it with a lock-in amplifier give consistent results, and can therefore be considered equivalent. For clarity, in the following we label as GDCG_{DC} the conductance extracted from VV-IbiasI_{bias} curves as in Figure 1e of the main text, GLIG_{LI} the conductance measured with a lock-in amplifier (as in Figure 3c to Figure 3f of the main text).

To directly compare GDCG_{DC} with GLIG_{LI} an additional alignment step between the maps is necessary. In fact, while IswI_{sw} and GDCG_{DC} are automatically aligned because the two quantities are extracted from the same dataset, this is not necessarily true for maps from different datasets, as it is the case GLIG_{LI} and GDCG_{DC}. Every thermal cycle and tip approach routine can lead to a small drift of the scan area with respect to previous SGM maps, making the superposition between different datasets not exact. As a figure of merit to evaluate the optimal alignment between the maps, we calculate the Pearson correlation coefficient using the scipy.stats.pearsonr Python function, and we search for the shift that maximizes the correlation coefficient. For two-dimensional maps, this parameter can be calculated by first unraveling each two-dimensional map into a one-dimensional ordered array.

The GDCG_{DC} and GLIG_{LI} maps have different scan size and resolution: the GDCG_{DC} map has pixels of 20×8020 \times80 nm over an area of 5×45 \times4 μ\mum2^2 along xtipx_{tip} and ytipy_{tip} directions, while the GLIG_{LI} map has pixels of 20×2020 \times20 nm, over an area of 4.5×44.5 \times4 μ\mum2^2. As a first step, we interpolate the GDCG_{DC} map in order to match the GLIG_{LI} map resolution, and crop it such that the range along the xtipx_{tip}-axis matches that of the smaller GLIG_{LI} map. The original GDCG_{DC} map is shown in Figure S2, the interpolated cropped map (corresponding to the area in Figure S2 limited by the vertical white line) is shown in Figure S2. The correlation analysis is done in the (0,4.5)(0, 4.5) μ\mum range along xtipx_{tip} (and full range along ytipy_{tip}). In Figure S2 we report the original GLIG_{LI} map, that has the same spatial extension of the GDCG_{DC} map in Figure S2. We next iterate the calculation of the Pearson correlation coefficient for varying spatial shifts, searching for the (xshift,yshift)(x_{shift}, y_{shift}) coordinates that maximize the correlation coefficient. The map of the correlation coefficient as a function of (xshift,yshift)(x_{shift}, y_{shift}) is shown in Figure S2. The (xshift,yshift)(x_{shift}, y_{shift}) coordinates indicate the spatial shift of the GDCG_{DC} map with respect to the GLIG_{LI} map. The optimal shift coordinates are (xshift,yshift)=(−500,−180)(x_{shift}, y_{shift}) = (-500, -180) nm. The resulting optimally shifted GDCG_{DC} map is shown in Figure S2.

Figure S2. (a) SGM map of junction conductance GDCG_{DC} obtained by fitting the dissipative branch of VV-IbiasI_{bias} curves (from the same dataset as in Figure 1e of the main text). VBG=−20V_{BG} = -20 V, Vtip=10V_{tip} = 10 V. (b) Interpolated SGM map of junction conductance GDCG_{DC}, such that pixels have size 20×2020 \times20 nm. (c) SGM map of junction conductance GLIG_{LI} measured with a lock-in amplifier (from the same dataset as in Figure 3f of the main text). (d) Pearson correlation coefficient between GLIG_{LI} and GDCG_{DC} maps as a function of the relative spatial shift between the two maps. The shift is the translation of the GDCG_{DC} map with respect to the GLIG_{LI} map. (e) Map of GDCG_{DC} shifted by the optimal shift coordinates extracted from panel d. (f) Comparison of linecuts of GLIG_{LI} (upper panel) and GDCG_{DC} (lower panel) at ytip=2y_{tip} = 2 μ\mum, showing that GLIG_{LI} and GDCG_{DC} share the same oscillatory pattern. The GDCG_{DC} linecut is obtained as the average of three lines around ytip=2y_{tip} = 2 μ\mum.

In Figure S2 we compare line cuts taken at ytip=2y_{tip} = 2 μ\mum. Analogously to line cuts of IswI_{sw} and GG shown in Figure 1f of the main text, local maxima and minima in GLIG_{LI} and GDCG_{DC} are aligned, as exemplified by the guiding black dashed lines, and demonstrate that it is equivalent to measure the junction conductance either with a lock-in amplifier or by extracting the conductance from VV-IbiasI_{bias} curves. This comparison also underlines the exceptional reproducibility and stability (both spatial and temporal) of the observed fringe patterns.

We finally identify the (2,2)(2, 2) μ\mum as the coordinates of the junction center (black dots in Figure S2 and Figure S2). SGM spatial maps in the main text (Figure 1c,Figure 1e and Figure 3c to Fig. 3f) are plotted such that the center is shifted to (0,0)(0, 0) coordinates, and cropped so that they are symmetric around this point. As a result, the IswI_{sw} and GDCG_{DC} maps of Figure 1c,Figure 1e extend in a 5×3.65 \times3.6 μ\mum2^2, while GLIG_{LI} maps shown in Figure 3c to Fig. 3f extend over a 4×3.64 \times3.6 μ\mum2^2 area.

Polynomial background removal for ΔG\Delta G

To isolate Fabry-Pérot oscillations in ΔG\Delta G shown in Figure 2c of the main text, a polynomial background is removed. This is obtained by fitting a polynomial surface to the raw conductance map G(VBG,Vtip)G(V_{BG}, V_{tip}) shown in Figure S3 using a linear least-squares fit. The resulting smooth background GbackgroundG_{\mathrm{background}} (Figure S3) is then subtracted from the measured data to yield the residual conductance oscillations, ΔG=G−Gbackground\Delta G = G - G_{\mathrm{background}}.

Figure S3. (a) Zoom of Figure 2b of the main text, in the VBG=(−20,−13)V_{BG} = (-20, -13) V range. (b) Polynomial background, obtained by performing a linear least-squares fit to the data in panel a.

Analytical model for the lensing effect

In this section we show the details of the analytical model that demonstrates the lensing effect. We extend the model described in [20], where authors consider translation invariance along the interfaces. Since the tip breaks this symmetry, we here performed a spatially resolved scattering analysis. Defining xx as the coordinate parallel to the cavity interfaces, one can consider scattering modes with well-defined xx (see Figure S2, Figure S5). For the Heisenberg uncertainty principle, these modes consist of a quantum superposition of momenta kxk_x. Since the interfaces, through the Klein collimation [20, 21], act as a filter for the quasi-normal modes kx≪kFk_x \ll k_F, and if the tip potential varies slowly with respect to kFk_F in the xx-direction, one can assume that the non-filtered part of the superposition in kxk_x is locally conserved, where locally means that the conservation is true at a fixed xx. The Dirac equation for the sublattice spinor ψ\psi of a scattering mode with well-defined xx then reads

Top-view sketch of the graphene junction and chemical potential profile

Figure S5. Panel a shows a top-view sketch of the graphene junction. Blue regions represent the superconducting contacts with Fermi energy μleads\mu_{\mathrm{leads}}. Green regions are the nn-doped parts of graphene with smoothly interpolating chemical potential μ(y)\mu(y). The yellow region is the graphene area with flat Fermi energy μp\mu_p, tunable via the backgate voltage. Panel b shows explicitly the spatial dependence of μ(y)\mu(y) along the junction.

i∂yψ=[(μp+Vtip(x,y))σz−ikxσx]ψ,(4)i\partial_y\psi= [(\mu_p + V_{\mathrm{tip}}(x,y))\sigma_z - ik_x\sigma_x]\psi, \tag*{(4)}

where μp\mu_p is the chemical potential of the pp-cavity, and σi\sigma_i are the Pauli matrices. Here, without loss of generality, we set the Fermi energy to zero. From the Dirac equation follows ky(x,y)=(μp+Vtip(x,y))2−kx2k_y(x,y) = \sqrt{(\mu_p + V_{\mathrm{tip}}(x,y))^2 - k_x^2}. The scattering amplitude Ar→rfA_{r\to r_f} from rr (at the first encountered interface) to rfr_f (on the opposite interface) can be written in a path-integral fashion

Ar→rf∝∫dkx2πeikx(xf−x)∑FP pathseiS[x,kx]/ℏ,(5)A_{r\to r_f} \propto\int\frac{d k_x}{2\pi}e^{ik_x(x_f-x)}\sum_{\mathrm{FP\ paths}}e^{iS[x,k_x]/\hbar}, \tag*{(5)}

where the sum is over the Fabry-Pérot paths and SS is the action [62]. This sum is a geometric series, and so the amplitude results in

Ar→rf∝∫dkx2πeikx(xf−x)t(kx,μp)eiθWKB(x,kx)1−r(kx,μp)e2iθWKB(x,kx),(6)A_{r\to r_f} \propto\int\frac{d k_x}{2\pi}e^{ik_x(x_f-x)}\frac{t(k_x,\mu_p)e^{i\theta_{\mathrm{WKB}}(x,k_x)}}{1-r(k_x,\mu_p)e^{2i\theta_{\mathrm{WKB}}(x,k_x)}}, \tag*{(6)}

where t=1−rt=1-r is the transmission probability across the interface, and

θWKB(x,kx)=∫−LcLcky(x,y,kx) dy.(7)\theta_{\mathrm{WKB}}(x,k_x)=\int_{-L_c}^{L_c}k_y(x,y,k_x)\,dy. \tag*{(7)}

Performing the approximation for Vtip(x,y)≃⟨Vtip⟩+vtip(x,y)V_{\mathrm{tip}}(x,y)\simeq\langle V_{\mathrm{tip}}\rangle+v_{\mathrm{tip}}(x,y) and a perturbative expansion in δVtip/(μp+⟨Vtip⟩)\delta V_{\mathrm{tip}}/(\mu_p+\langle V_{\mathrm{tip}}\rangle), the WKB angle [63] reduces to θWKB=θeff+δθ(x,kx)\theta_{\mathrm{WKB}}=\theta_{\mathrm{eff}}+\delta\theta(x,k_x). The conductance is

G∝∫−W/2W/2dx dxf∣Ar→rf∣2,(8)G\propto\int_{-W/2}^{W/2}dx\,dx_f\left|A_{r\to r_f}\right|^2, \tag*{(8)}

and this leads to

G=4e2h∫dkx2π∫−W/2W/2dx 11+Feff(kx,kF,eff)sin⁡2[kF,effLc+δθ(x−xtip)].(9)G=\frac{4e^2}{h}\int\frac{d k_x}{2\pi}\int_{-W/2}^{W/2}dx\,\frac{1}{1+F_{\mathrm{eff}}(k_x,k_{F,\mathrm{eff}})\sin^2[k_{F,\mathrm{eff}}L_c+\delta\theta(x-x_{\mathrm{tip}})]}. \tag*{(9)}

The explicit expression for δθ(x−xtip)∣ytip=0\delta\theta(x-x_{\mathrm{tip}})|_{y_{\mathrm{tip}}=0} is

δθ(x−xtip)=2(ηVtip)kF,effR2R2+(x−xtip)2arctan⁡(L2R2+(x−xtip)2).(10)\delta\theta(x-x_{\mathrm{tip}})=2(\eta V_{\mathrm{tip}})k_{F,\mathrm{eff}}\frac{R^2}{\sqrt{R^2+(x-x_{\mathrm{tip}})^2}}\arctan\left(\frac{L}{2\sqrt{R^2+(x-x_{\mathrm{tip}})^2}}\right). \tag*{(10)}

In Figure 2g of the main text we have shown how this formula predicts the conductance oscillations at a fixed effective Fermi momentum. In addition to the main harmonic, higher frequency oscillations are observed, which arise from the xx-integral over the transmission amplitudes T(x,kx)T(x,k_x) at a fixed kxk_x. Indeed, the xx-dependence of the function T(x,kx)=(1+Feff(kx,kF,eff)sin⁡2[kF,effLc+δθ(x−xtip)])−1T(x,k_x)=(1+F_{\mathrm{eff}}(k_x,k_{F,\mathrm{eff}})\sin^2[k_{F,\mathrm{eff}}L_c+\delta\theta(x-x_{\mathrm{tip}})])^{-1} exhibit a spatial pattern that reflects how a single mode feels the local potential of the tip modifying its phase. Integrating over the different patterns for each kxk_x results in the generation of the higher harmonics in the conductance. Here, LcL_c represents an effective length over which electrons start to feel the effect of the tip.

To further demonstrate the role of the tip in the lensing effect, we show how the conductance modulation depends on the Lorentzian radius. In Figure S4 we show, as a function of the tip radius RR, the conductance G(Vtip)G(V_{\mathrm{tip}}) after the polynomial background removal for a fixed kF,effk_{F,\mathrm{eff}}. As the radius is lowered, the main harmonic increases its period, reflecting the fact that less scattering modes have an impact parameter x−xtip≲Rx-x_{\mathrm{tip}}\lesssim R that contribute to the main interference effect.

Comparison of conductance oscillations for different tip radius

Figure S4. Comparison of conductance oscillations for different tip radius RR. Lowering the radius means increasing the period of the main harmonic, showing that the oscillations arise from the lensing of a sufficiently large number of incident modes with impact parameter x−xtip≲Rx-x_{\mathrm{tip}} \lesssim R. A comparison with the experimental curve is shown (light blue line).

Details on the numerical model

We describe here the theoretical framework underlying the simulations presented in the main text. We consider a planar junction lying in the xx-yy plane, with yy as the longitudinal coordinate parallel to the current flow. In Figure S5 we show a sketch of the simulations setup. To avoid possible edge bound states, not present in the experiment, we have chosen armchair edges along xx for the model of graphene. Its width is W×aW \times a along xx and its length L×aL \times a along yy, where aa is the coarse-grained lattice constant [49, 50, 57, 64]. The niobium interfaces are located at fixed values of yy and extend along xx. The tight-binding Hamiltonian is defined following Ref. [49, 50]. Because of the coarse-graining of aa, the hopping amplitude tt must be renormalized [57, 64] so that the graphene Fermi velocity vF∝atv_F \propto at is kept constant.

To model the Nb metallic contacts, we introduce a spatially modulated chemical potential μ=μ(y)\mu=\mu(y), with

μ(y<−L/2×a)=μ(y>L/2×a)≡μleads.\mu(y < -L/2 \times a) = \mu(y > L/2 \times a) \equiv\mu_{\mathrm{leads}}.

To account for the doping induced by the Nb contacts, we define a central region −l/2×a<y<l/2×a-l/2 \times a < y < l/2 \times a with a flat chemical potential μp\mu_p, representing the backgate-tunable Fermi energy of the graphene sheet. At each interface with the leads, μ(y)\mu(y) interpolates smoothly between μleads\mu_{\mathrm{leads}} and μp\mu_p over a region of length m×am \times a, with L=2m+lL=2m+l. These transition regions correspond to the nn-doped parts of graphene induced by proximity to the Nb contacts. Figure S5 shows a sketch of the junction (panel a) and the corresponding μ(y)\mu(y) profile (panel b).

Disorder is modeled as a sum of Gaussian puddles,

μdis(x,y)=∑n=1Nvnexp⁡(−(x−xn)2+(y−yn)2ξ2),\mu_{\mathrm{dis}}(x,y)=\sum_{n=1}^{N}v_n\exp\left(-\frac{(x-x_n)^2+(y-y_n)^2}{\xi^2}\right),

where each amplitude vnv_n is drawn from a Gaussian distribution with standard deviation VdisV_{\mathrm{dis}}, each puddle has correlation length ξ\xi, and the centers (xn,yn)(x_n,y_n) are uniformly distributed at random over the graphene region. Their number NN is set by requiring Nπξ2/WL≡fN\pi\xi^2/WL \equiv f, i.e., a fixed areal coverage fraction ff.

The SGM tip is modeled as a local modification of the Fermi energy background. Following Refs. [33, 49, 50, 58], the tip potential takes the Lorentzian form

μtip(x,y)=μ0,tip1+x2+y2R2(11)\mu_{\mathrm{tip}}(x,y)=\frac{\mu_{0,\mathrm{tip}}}{1+\frac{x^2+y^2}{R^2}} \tag*{(11)}

where RLorR_{\mathrm{Lor}} is the effective tip radius. To relate μp\mu_p and μ0,tip\mu_{0,\mathrm{tip}} to the experimental gate voltages, we note that in graphene the charge density n=μ2/[π(ℏvF)2]n=\mu^2/[\pi(\hbar v_F)^2] is a linear function of the gate voltage, which yields dμp2/dVbg=constd\mu_p^2/dV_{\mathrm{bg}}=\mathrm{const} and hence

μp(Vbg)=α sgn⁡(Vbg)∣Vbg∣,\mu_p(V_{\mathrm{bg}})=\alpha\,\operatorname{sgn}(V_{\mathrm{bg}})\sqrt{\lvert V_{\mathrm{bg}}\rvert},
μ0,tip(Vtip)=β sgn⁡(Vtip)∣Vtip∣.\mu_{0,\mathrm{tip}}(V_{\mathrm{tip}})=\beta\,\operatorname{sgn}(V_{\mathrm{tip}})\sqrt{\lvert V_{\mathrm{tip}}\rvert}.

These relations fully specify the mapping between the simulation parameters and the experimental control knobs. The conductance is computed in the linear-response regime via the Landauer formula

G=e2hTr⁡tt†,G=\frac{e^2}{h}\operatorname{Tr}tt^\dagger,

where tt is the transmission block of the scattering matrix between the two leads. All simulations are performed using Kwant [65]. The parameter values are listed in Table S1.

ParameterSymbolValue
Primitive vector lengthaa14 nm
Hopping energytt67.5 meV
Sample widthWW3 μ\mum
Sample lengthLL600 nm
Normal region lengthll440 nm
Interface lengthmm80 nm
Fermi energy of leadsμleads\mu_{\mathrm{leads}}64 meV
Conversion μp−Vbg\mu_p - V_{\mathrm{bg}}α\alpha27 meV ×\times V−1/2^{-1/2}
Conversion μ0,tip−Vtip\mu_{0,\mathrm{tip}} - V_{\mathrm{tip}}β\beta21.6 meV ×\times V−1/2^{-1/2}
Tip effective radiusRR436 nm
Standard deviation of disorder strengthVdisV_{\mathrm{dis}}1.35 meV
Disorder correlation lengthξ\xi60 nm
Disorder filling fractionff10%

Table S1. Summary of simulation parameters.

Acknowledgments

We acknowledge Simone Traverso for useful discussions, and support from projects PRIN2022 2022-PH852L-(PE3) TopoFlag and PNRR MUR Project No. PE0000023-NQSTI. K. W. and T. T. acknowledge support from the CREST (JPMJCR24A5), JST and World Premier International Research Center Initiative (WPI), MEXT, Japan.

References

  1. [1]Chen, S.; Han, Z.; Elahi, M. M.; Habib, K. M. M.; Wang, L.; Wen, B.; Gao, Y.; Taniguchi, T.; Watanabe, K.; Hone, J.; Ghosh, A. W.; Dean, C. R. Electron optics with p-n junctions in ballistic graphene. Science 2016, 353, 1522–1525.pubmed.ncbi.nlm.nih.gov/27708099
  2. [2]Chakraborti, H.; Gorini, C.; Knothe, A.; Liu, M.-H.; Makk, P.; Parmentier, F. D.; Perconte, D.; Richter, K.; Roulleau, P.; Sacépé, B.; Schönenberger, C.; Yang, W. Electron wave and quantum optics in graphene. Journal of Physics: Condensed Matter 2024, 36, 393001.DOI
  3. [3]Ji, Y. Y.; Chung, Y.; Sprinzak, D.; Heiblum, M.; Mahalu, D.; Shtrikman, H. An electronic Mach–Zehnder interferometer. Nature 2003, 422, 415–418.pubmed.ncbi.nlm.nih.gov/12660779
  4. [4]Bocquillon, E.; Freulon, V.; Parmentier, F. D.; Berroir, J.-M.; Plaçais, B.; Wahl, C.; Rech, J.; Jonckheere, T.; Martin, T.; Grenier, C.; Ferraro, D.; Degiovanni, P.; Fève, G. Electron quantum optics in ballistic chiral conductors. Annalen der Physik 2014, 526, 1–30.
  5. [5]Acciai, M.; Ronetti, F.; Ferraro, D.; Rech, J.; Jonckheere, T.; Sassetti, M.; Martin, T. Levitons in superconducting point contacts. Phys. Rev. B 2019, 100, 085418.DOI
  6. [6]Bartolomei, H.; Kumar, M.; Bisognin, R.; Marguerite, A.; Berroir, J.-M.; Bocquillon, E.; Plaçais, B.; Cavanna, A.; Dong, Q.; Gennser, U.; Jin, Y.; Fève, G. Fractional statistics in anyon collisions. Science 2020, 368, 173–177.DOI
  7. [7]Nakamura, J.; Liang, S.; Gardner, G. C.; Manfra, M. J. Direct observation of anyonic braiding statistics. Nature Physics 2020, 16, 931–936.DOI
  8. [8]Carrega, M.; Chiroli, L.; Heun, S.; Sorba, L. Anyons in quantum Hall interferometry. Nature Reviews Physics 2021, 3, 698–711.DOI
  9. [9]Ronetti, F.; Bertin-Johannet, B.; Popoff, A.; Rech, J.; Jonckheere, T.; Grémaud, B.; Raymond, L.; Martin, T. Levitons in correlated nano-scale systems. Chaos: An Interdisciplinary Journal of Nonlinear Science 2024, 34, 042103.DOI
  10. [10]Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Katsnelson, M. I.; Grigorieva, I. V.; Dubonos, S. V.; Firsov, A. A. Two-dimensional gas of massless Dirac fermions in graphene. Nature 2005, 438, 197–200.DOI
  11. [11]Zhang, Y.; Tan, Y.-W.; Stormer, H. L.; Kim, P. Experimental observation of the quantum Hall effect and Berry’s phase in graphene. Nature 2005, 438, 201–204.pubmed.ncbi.nlm.nih.gov/16281031
  12. [12]Katsnelson, M. I.; Novoselov, K. S.; Geim, A. K. Chiral tunnelling and the Klein paradox in graphene. Nature Physics 2006, 2, 620–625.DOI
  13. [13]Allain, P. E.; Fuchs, J. N. Klein tunnelling in graphene: optics with massless electrons. The European Physical Journal B 2011, 83, 301–317.DOI
  14. [14]Lee, G.-H.; Park, G.-H.; Lee, H.-J. Observation of negative refraction of Dirac fermions in graphene. Nature Physics 2015, 11, 925–929.DOI
  15. [15]Cheianov, V. V.; Fal’ko, V.; Altshuler, B. L. The Focusing of Electron Flow and a Veselago Lens in Graphene p − n Junctions. Science 2007, 315, 1252–1255.pubmed.ncbi.nlm.nih.gov/17332407
  16. [16]Williams, J. R.; Low, T.; Lundstrom, M. S.; Marcus, C. M. Gate-controlled guiding of electrons in graphene. Nature Nanotechnology 2011, 6, 222–225.DOI
  17. [17]Zhang, X.; Ren, W.; Bell, E.; Zhu, Z.; Tsai, K.-T.; Luo, Y.; Watanabe, K.; Taniguchi, T.; Kaxiras, E.; Luskin, M.; Wang, K. Gate-tunable Veselago interference in a bipolar graphene microcavity. Nature Communications 2022, 13, 6711.DOI
  18. [18]Déprez, C.; Veyrat, L.; Vignaud, H.; Nayak, G.; Watanabe, K.; Taniguchi, T.; Gay, F.; Sellier, H.; Sacépé, B. A tunable Fabry–Pérot quantum Hall interferometer in graphene. Nature Nanotechnology 2021, 16, 555–562.DOI
  19. [19]Kim, J.; Dev, H.; Kumar, R.; Ilin, A.; Haug, A.; Bhardwaj, V.; Hong, C.; Watanabe, K.; Taniguchi, T.; Stern, A.; Ronen, Y. Aharonov–Bohm interference and statistical phase-jump evolution in fractional quantum Hall states in bilayer graphene. Nature Nanotechnology 2024, 19, 1619–1626.DOI
  20. [20]Shytov, A. V.; Rudner, M. S.; Levitov, L. S. Klein Backscattering and Fabry-Pérot Interference in Graphene Heterojunctions. Phys. Rev. Lett. 2008, 101, 156804.DOI
  21. [21]Young, A. F.; Kim, P. Quantum interference and Klein tunnelling in graphene heterojunctions. Nature Physics 2009, 5, 222–226.DOI
  22. [22]Nam, S.-G.; Ki, D.-K.; Park, J. W.; Kim, Y.; Kim, J. S.; Lee, H.-J. Ballistic transport of graphene pnp junctions with embedded local gates. Nanotechnology 2011, 22, 415203.DOI
  23. [23]Topinka, M. A.; LeRoy, B. J.; Westervelt, R. M.; Shaw, S. E. J.; Fleischmann, R.; Heller, E. J.; Maranowski, K. D.; Gossard, A. C. Coherent branched flow in a two-dimensional electron gas. Nature 2001, 410, 183–186.
  24. [24]Crook, R.; Graham, A. C.; Smith, C. G.; Farrer, I.; Beere, H. E.; Ritchie, D. A. Erasable electrostatic lithography for quantum components. Nature 2003, 424, 751–754.
  25. [25]Da Cunha, C. R.; Aoki, N.; Morimoto, T.; Ochiai, Y.; Akis, R.; Ferry, D. K. Imaging of quantum interference patterns within a quantum point contact. Applied Physics Letters 2006, 89, 242109.
  26. [26]Aidala, K. E.; Parrott, R. E.; Kramer, T.; Heller, E. J.; Westervelt, R. M.; Hanson, M. P.; Gossard, A. C. Imaging magnetic focusing of coherent electron waves. Nature Physics 2007, 3, 464–468.DOI
  27. [27]Martins, F.; Hackens, B.; Pala, M. G.; Ouisse, T.; Sellier, H.; Wallart, X.; Bollaert, S.; Cappy, A.; Chevrier, J.; Bayot, V.; Huant, S. Imaging Electron Wave Functions Inside Open Quantum Rings. Phys. Rev. Lett. 2007, 99, 136807.DOI
  28. [28]Paradiso, N.; Heun, S.; Roddaro, S.; Sorba, L.; Beltram, F.; Biasiol, G.; Pfeiffer, L. N.; West, K. W. Imaging Fractional Incompressible Stripes in Integer Quantum Hall Systems. Phys. Rev. Lett. 2012, 108, 246801.DOI
  29. [29]Toussaint, S.; Martins, F.; Faniel, S.; Pala, M. G.; Desplangue, L.; Wallart, X.; Sellier, H.; Huant, S.; Bayot, V.; Hackens, B. On the origins of transport inefficiencies in mesoscopic networks. Scientific Reports 2018, 8, 3017.DOI
  30. [30]Connolly, M. R.; Chiu, K. L.; Lombardo, A.; Fasoli, A.; Ferrari, A. C.; Anderson, D.; Jones, G. A. C.; Smith, C. G. Tilted potential induced coupling of localized states in a graphene nanoconstriction. Phys. Rev. B 2011, 83, 115441.DOI
  31. [31]Connolly, M. R.; Herbschleb, E. D.; Puddy, R. K.; Roy, M.; Anderson, D.; Jones, G. A. C.; Maksym, P.; Smith, C. G. Reading and writing charge on graphene devices. Applied Physics Letters 2012, 101, 023505.DOI
  32. [32]Garcia, A. G. F.; König, M.; Goldhaber-Gordon, D.; Todd, K. Scanning gate microscopy of localized states in wide graphene constrictions. Phys. Rev. B 2013, 87, 085446.DOI
  33. [33]Herbschleb, E. D.; Puddy, R. K.; Marconcini, P.; Griffiths, J. P.; Jones, G. A. C.; Macucci, M.; Smith, C. G.; Connolly, M. R. Direct imaging of coherent quantum transport in graphene p−n−p junctions. Phys. Rev. B 2015, 92, 125414.DOI
  34. [34]Bours, L.; Guiducci, S.; Mreńca-Kolaśińska, A.; Szafran, B.; Maan, J. C.; Heun, S. Manipulating quantum Hall edge channels in graphene through scanning gate microscopy. Phys. Rev. B 2017, 96, 195423.DOI
  35. [35]Cabosart, D.; Felten, A.; Reckinger, N.; Iordanescu, A.; Toussaint, S.; Faniel, S.; Hackens, B. Recurrent Quantum Scars in a Mesoscopic Graphene Ring. Nano Letters 2017, 17, 1344–1349.DOI
  36. [36]Brun, B.; Moreau, N.; Somanچی, S.; Nguyen, V.-H.; Watanabe, K.; Taniguchi, T.; Charlier, J.-C.; Stampfer, C.; Hackens, B. Imaging Dirac fermions how through a circular Veselago lens. Phys. Rev. B 2019, 100, 041401(R).DOI
  37. [37]Marguerite, A.; Birkbeck, J.; Aharon-Steinberg, A.; Halbertal, D.; Bagani, K.; Marcus, I.; Myasoedov, Y.; Geim, A. K.; Perello, D. J.; Zeldov, E. Imaging work and dissipation in the quantum Hall state in graphene. Nature 2019, 575, 628–633.
  38. [38]Brun, B.; Nguyen, V.-H.; Moreau, N.; Somanچی, S.; Watanabe, K.; Taniguchi, T.; Charlier, J.-C.; Stampfer, C.; Hackens, B. Graphene Whisperitronics: Transducing Whispering Gallery Modes into Electronic Transport. Nano Letters 2022, 22, 128–134.DOI
  39. [39]Gold, C.; Knothe, A.; Kurzmann, A.; Garcia-Ruiz, A.; Watanabe, K.; Taniguchi, T.; Fal’ko, V.; Ensslin, K.; Ihn, T. Coherent Jetting from a Gate-Defined Channel in Bilayer Graphene. Phys. Rev. Lett. 2021, 127, 046801.DOI
  40. [40]Calado, V. E.; Goswami, S.; Nanda, G.; Diez, M.; Akhmerov, A. R.; Watanabe, K.; Taniguchi, T.; Klapwijk, T. M.; Vandersypen, L. M. K. Ballistic Josephson junctions in edge-contacted graphene. Nature Nanotechnology 2015, 10, 761–764.DOI
  41. [41]Ben Shalom, M.; Zhu, M. J.; Fal’ko, V. I.; Mishchenko, A.; Kretinin, A. V.; Novoselov, K. S.; Woods, C. R.; Watanabe, K.; Taniguchi, T.; Geim, A. K.; Prance, J. R. Quantum oscillations of the critical current and high-field superconducting proximity in ballistic graphene. Nature Physics 2016, 12, 318–322.DOI
  42. [42]Allen, M. T.; Shtanko, O.; Fulga, I. C.; Wang, J. I.-J.; Nurgaliev, D.; Watanabe, K.; Taniguchi, T.; Akhmerov, A. R.; Jarillo-Herrero, P.; Levitov, L. S.; Yacoby, A. Observation of Electron Coherence and Fabry–Perot Standing Waves at a Graphene Edge. Nano Letters 2017, 17, 7380–7386.DOI
  43. [43]Villani, I.; Carrega, M.; Crippa, A.; Strambini, E.; Giazzotto, F.; Mišeikis, V.; Coletti, C.; Beltram, F.; Watanabe, K.; Taniguchi, T.; Heun, S.; Pezzini, S. Quasi-Φ₀-Periodic Supercurrent at Quantum Hall Transitions. ACS Nano 2025, 19, 27370–27378.DOI
  44. [44]Kaperek, K.; Heun, S.; Carrega, M.; Wójcik, P.; Nowak, M. P. Theory of scanning gate microscopy imaging of the supercurrent distribution in a planar Josephson junction. Phys. Rev. B 2022, 106, 035432.DOI
  45. [45]Maji, S.; Sowa, K.; Nowak, M. P. Scanning gate microscopy of nontracing electron-hole trajectories in a normal-superconductor junction. Phys. Rev. B 2024, 109, 115410.DOI
  46. [46]Maji, S.; Nowak, M. P. Aharonov-Bohm interference probing of chiral Andreev edge states via scanning gate microscopy, arXiv:2609.27716 [cond-mat.mes-hall], 2026.arxiv.org/abs/2609.27716
  47. [47]Hegedüs, M.; Banerjee, R.; Hutcheson, A.; Barker, T.; Mahashabde, S.; Danilov, A. V.; Kubatkin, S. E.; Antonov, V.; de Graaf, S. E. In situ scanning gate imaging of individual quantum two-level system defects in live superconducting circuits. Science Advances 2025, 11, eadt8586.DOI
  48. [48]Bhandari, S.; Lee, G.-H.; Watanabe, K.; Taniguchi, T.; Kim, P.; Westervelt, R. M. Imaging Andreev Reflection in Graphene. Nano Letters 2020, 20, 4890–4894.DOI
  49. [49]Lombardi, A.; Shukla, G.; Bucci, G.; Salimian, S.; Zannier, V.; Traverso, S.; Fracassi, S.; Traverso Ziani, N.; Sassetti, M.; Carrega, M.; Beltram, F.; Sorba, L.; Heun, S. Supercurrent modulation in InSb nanoflag-based Josephson junctions by scanning gate microscopy. Communications Materials 2025, 6, 272.DOI
  50. [50]Villani, I.; Fracassi, S.; Traverso Ziani, N.; Sassetti, M.; Carrega, M.; Mišeikis, V.; Coletti, C.; Beltram, F.; Watanabe, K.; Taniguchi, T.; Pezzini, S.; Heun, S. Scanning gate microscopy modulation of supercurrent in graphene Josephson junctions. Applied Physics Letters 2026, 129, 162601.
  51. [51]Miseikis, V.; Convertino, D.; Mishra, N.; Gemmi, M.; Mashoff, T.; Heun, S.; Haghighian, N.; Bisio, F.; Canepa, M.; Piazza, V.; Coletti, C. Rapid CVD growth of millimetre-sized single crystal graphene using a cold-wall reactor. 2D Materials 2015, 2, 014006.DOI
  52. [52]Pezzini, S.; Mišeikis, V.; Pace, S.; Rossella, F.; Watanabe, K.; Taniguchi, T.; Coletti, C. High-quality electrical transport using scalable CVD graphene. 2D Materials 2020, 7, 041003.DOI
  53. [53]Villani, I.; Chiroli, L.; Carrega, M.; Crippa, A.; Strambini, E.; Giazzotto, F.; Mišeikis, V.; Coletti, C.; Beltram, F.; Watanabe, K.; Taniguchi, T.; Heun, S.; Pezzini, S. Graphene Josephson diodes from inherent asymmetric disorder, arXiv:2606.17673 [cond-mat.supr-con], 2026.arxiv.org/abs/2606.17673
  54. [54]Tinkham, M., Introduction to Superconductivity, 2nd Ed.; McGraw-Hill: New York, 1996.
  55. [55]Borzenets, I. V.; Amet, F.; Ke, C. T.; Draelos, A. W.; Wei, M. T.; Seredinski, A.; Watanabe, K.; Taniguchi, T.; Bomze, Y.; Yamamoto, M.; Tarucha, S.; Finkelstein, G. Ballistic Graphene Josephson Junctions from the Short to the Long Junction Regimes. Phys. Rev. Lett. 2016, 117, 237002.
  56. [56]Born, M.; Wolf, E., Principles of Optics, Electromagnetic Theory of Propagation, Interference and Diffraction of Light, 7th ed.; Cambridge University Press: Cambridge, 1999.
  57. [57]Liu, M.-H.; Rickhaus, P.; Makk, P.; Tóvári, E.; Maurand, R.; Tkatschenko, F.; Weiss, M.; Schönemberger, C.; Richter, K. Scalable Tight-Binding Model for Graphene. Phys. Rev. Lett. 2015, 114, 036601.pubmed.ncbi.nlm.nih.gov/25659011
  58. [58]Szafran, B. Scanning gate microscopy simulations for quantum rings: Effective potential of the tip and conductance maps. Phys. Rev. B 2011, 84, 075336.DOI
  59. [59]Wang, L.; Meric, I.; Huang, P. Y.; Gao, Q.; Gao, Y.; Tran, H.; Taniguchi, T.; Watanabe, K.; Campos, L. M.; Muller, D. A.; Guo, J.; Kim, P.; Hone, J.; Shepard, K. L.; Dean, C. R. One-Dimensional Electrical Contact to a Two-Dimensional Material. Science 2013, 342, 614–617.DOI
  60. [60]Purdie, D. G.; Pugno, N. M.; Taniguchi, T.; Watanabe, K.; Ferrari, A. C.; Lombardo, A. Cleaning interfaces in layered materials heterostructures. Nat. Commun. 2018, 9, 5387.DOI
  61. [61]Miseikis, V.; Bianco, F.; David, J.; Gemmi, M.; Pellegrini, V.; Romagnoli, M.; Coletti, C. Deterministic patterned growth of high-mobility large-crystal graphene: a path towards wafer scale integration. 2D Mater. 2017, 4, 021004.arxiv.org/abs/1611.00923
  62. [62]Feynman, R. P. Space-Time Approach to Non-Relativistic Quantum Mechanics. Rev. Mod. Phys. 1948, 20, 367–387.
  63. [63]Miller, S. C.; Good, R. H. A WKB-Type Approximation to the Schrödinger Equation. Phys. Rev. 1953, 91, 174–179.DOI
  64. [64]Liu, M.-H.; De Beule, C.; Mrenca-Kolасinska, A.; Wu, H.-Y.; Garcia-Ruiz, A.; Kochan, D.; Richter, K. Scalable tight-binding model for strained graphene. Phys. Rev. B 2026, 113, 195429.DOI
  65. [65]Groth, C. W.; Wimmer, M.; Akhmerov, A. R.; Waintal, X. Kwant: a software package for quantum transport. New J. Phys. 2014, 16, 063065.

Paper details

vlnc:W_8mf32g58
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents