Arxiv:1611.00227V4 [Quant-Ph] 6 Mar 2017 Tomechanics As the Mechanical Element in the Form of a Mechanical Systems (NEMS) [38]

Nonlinear Graphene Quantum Capacitors for Electro-optics

Sina Khorasani School of Electrical Engineering, Sharif University of Technology, Tehran, Iran and Ecole´ Polytechnique F´ed´erale de Lausanne, CH-1015, Lausanne, Switzerland∗

Akshay Koottandavida UM-DAE Centre for Excellence in Basic Sciences, Vidyanagari, Mumbai, 400098, India and Ecole´ Polytechnique F´ed´erale de Lausanne, CH-1015, Lausanne, Switzerland∗

Owing to its peculiar energy dispersion, the quantum capacitance property of graphene can be exploited in a two-dimensional layered capacitor configuration. Using graphene and boron nitride respectively as the electrodes and the insulating dielectric, a strongly nonlinear behavior at zero bias and small voltages is obtained. When the temperature is sufficiently low, the strong nonlinear interaction emerging from the quantum capacitance exhibits a diverse range of phenomena. The proposed structure could take over the functionalities of nonlinear elements in many cryogenic quantum systems, and in particular, quantum electro-optics. It is shown that ultrastrong coupling is easily reached with small number of pump photons at temperatures around 1K and capacitor areas of the order of 1µm2. A measure of anharmonicity is defined and as potential applications, a qubit design as well as schemes for non-reciprocal devices such as an electromagnetic frequency circulator are discussed.

INTRODUCTION ternative design for non-dissipative microwave quantum electro-optics [26–28]. In the fields of quantum optomechanics [1–5] and superconducting circuits, it is desirable to have a nonlinear It is required that the Boron nitride (BN), or any other interaction between modes with different frequencies. For substitute wide-band-gap 2D material, be at least a few this purpose, optomechanical crystals [1, 2, 6], nanome- monolayers thick to prevent tunnel currents. This is to chanical capacitors [2, 3, 7], kinetic inductors [8, 9], and avoid degrading the performance of the quantum capac- Josephson junctions [10–14] are being explored. By con- itor. The total electrical property of this layered struc- trast, in superconductive quantum computing [15–17], ture is then determined by a geometric parallel-plate ca- circuit analogues of quadratic optomechanics [18], and pacitor, placed in series with the graphene quantum ca- more recently realization of meminductors and memca- pacitor. The quantum capacitance at zero bias and low pacitors [19] have been proposed. temperatures takes on very small values, which overrides While nanomechanical capacitors effectively appear the geometric capacitance. Hence, the nonlinear quan- as third-order nonlinear Kerr elements in the circuits, tum capacitance would dominate the overall behavior at Josephson junctions, depending on the structure fabri- sufficiently low temperatures and zero or small bias volt- cation type and magnetic flux bias, may exhibit various ages. Electronic properties of graphene [29, 30], stem- degrees of second or third order nonlinear flux/current ming from its Dirac-cone shaped band structure are be- behavior [20]. Josephson junctions, despite their strong hind this quantum capacitance effect [31, 32]. In fact, the nonlinearity, normally operate at very small currents, nonlinear property of this quantum effect has not been which limits their maximum total photon occupation unnoticed and used by other authors for different applica- number. Moreover, they occasionally need to be biased tions such as sensors [33, 34]. Quantum capacitance has by either a DC magnetic field or electric current. As for been observed in Cooper-pair transistors [35] and boxes the nonlinear capacitors, Graphene has been used in op- [36], as well as qubit resonantor [37] and Nano-Electro- arXiv:1611.00227v4 [quant-ph] 6 Mar 2017 tomechanics as the mechanical element in the form of a Mechanical Systems (NEMS) [38]. drum capacitor [21–23]. Here, we suggest an alternative solution by proposing the design of a nonlinear quantum capacitance. The ca- It is shown that the strongly nonlinear feature of quan- pacitor is formed by a graphene-boron nitride-graphene tum capacitance may be well exploited to obtain an in- layered two-dimensional (2D) structure. This stacked 2D teraction Hamiltonian between multiple electromagnetic structure has been extensively studied for its various elec- modes, with an adjustable interaction rate through electronic and optoelectronic properties [24, 25]. This par- trostatic gating, all the way down to the temperatures ticular design avoids physical mechanical motion as the around zero. The existence of such a nonlinear capac- structure is not suspended. In fact, this provides an al- itance element compatible with low temperatures and quantum regimes, may significantly contribute to the design flexibility and the application range of superconducting circuit [15–17], circuit quantum cavity eletrodynam- ∗ [email protected]; sina.khorasani@epfl.ch ics, as well as quantum circuits and networks [39]. 2

I. THEORY

Quantum Capacitance

The per-unit capacitance of the 2D layered structure shown in Fig. 1 is composed of two components, which arise from geometric and quantum properties. These two appear in series as C = CG||CQ, where CG = /t is the geometric capacitance per unit area with and t being the permittivity and thickness of the BN layer. Also, d CQ = dV Q, is the quantum capacitance per unit area of the graphene electrode pair given by [40–42] 2 2e kBT EF FIG. 1. Schematics of the proposed layered structure. CQ = 2 ln 2 1 + cosh , (1) π(~vF ) kBT √ where e is the electronic charge, kBT is the thermal en- 1 3 where U = 2π v sgn(Q)N 2 with |Q| = eN is the ergy, v ≈ c/300 is the Fermi velocity of graphene, and 3 ~ F F total stored energy per unit area. These relations, despite E = eV/2 is the Fermi energy at either of the graphene F their simplicity, are unfortunately both singular at V = electrodes, with V being the voltage drop across C . In Q 0 due to the appearance of absolute value | · |, which general, V is not the same as the external applied volt- prohibits further expansion. Hence, (3) cannot be used age V , but under condition C >> C discussed be- ext G Q as a basis of field quantization. low, it is reasonable to ignore this difference and thus let However, it is possible to proceed directly from (1) V ≈ V . It should be however mentioned that achiev- ext at finite temperature instead, which after considerable ing this regime might be not quite trivial due to disorder algebra, allows us to obtain the series expansions and charge impurities which induce electron-hole puddles [43]. Furthermore, at very small E , the Fermi velocity 2ek T e2 F N ≈ B ln(2)V + V 3 + ... , (4) v in graphene should be renormalized due to electron- 2 2 F π(~vF ) 96(kBT ) electron interactions, leading to further extra logarithmic 2 ( 2 2 4 ) corrections [44]. π(~vF ) N π ~vF 4 −2 U ≈ − N + ... . The typical values of CQ is in the range of 1−10fFµm 2kBT ln(16) 4 ln(16)kBT [42]. In the ultralow temperature limit, the above expression approaches the linear form Firstly, it is seen from the first expansion term of the second equation, that the linear capacitance per unit area e3 lim C = |V |. (2) is given as T →0 Q π( v )2 ~ F 2 2e kBT ln(16) Hence, as the extreme limit for the symmetric structure C0 = 2 . (5) π(~vF ) under consideration with EF = 0 at zero bias and zero temperature, the differential capacitance vanishes CQ = This is to be compared with the geometrical capacitance 0. Though not being practical, this limit underlines the CG. The permittivity of BN is = 4.00 [24]. Hence tak- fact that the overall capacitance of the structure at zero ing the dielectric thickness to be t = 7nm, being sufficient bias, sufficiently low temperatures, and low drive-voltage to stop tunneling current, gives a geometric capacitance −2 is solely determined by its quantum properties. Although value of CG = 5.06fFµm . At the typical tempera- −2 being a bit counter-intuitive, this is exactly what one ture of 1K, we have C0 = 0.0563fFµm , which is much would need to exploit the resulting nonlinear property smaller than the geometric capacitance by two orders of of the quantum capacitor. In what follows, we explore magnitude. Hence, our initial assertion to ignore the ef- the limits of this dominance, and calculate the expected fect of the geometric capacitance at low temperatures and nonlinear interactions in details. zero bias is satisfied to a high accuracy. This situation is illustrated in Fig. 2. In general, a dielectric thickness of no less than 2nm should be sufficient to completely Series Expansion disallow tunneling. It could be made thick as long as CQ dominates by an order of magnitude, yielding the avail- At zero temperature, following (2), we obtain the non- able range of dielectric thickness as 3nm < t < 70nm. linear charge-voltage relationships Since, the monolayer BN is around only 2.489A˚ in thick- Z V e3 ness [24], t is sufficiently high to warrant an expitaxial Q = CdV = 2 |V |V, (3) growth. Structures of comparable type are being fabri- 0 2π(~vF ) cated in integrated optics [45, 46]. In general, there is no Z V 3 e 3 reason why other crystalline dielectrics such as SiO2 or U = QdV = 2 |V | . 0 6π(~vF ) Al2O3 could not be used. 3

5 Here, S is the total capacitor area. After adding a res- 1 2 ) onant inductive term to the Hamiltonian as Lqˆ˙ with 2 Room Temperature 2 qˆ = SQˆ being the operator of total charge on the 2D 1 √ quantum capacitor and ω = 1/ SLC0, this allows us to 0.50 T=30K indicate the Hamiltonian of such a sinusoidal ﬁeld from T=10K (4) as 0.10 1 τ Hˆ = ω aˆ†aˆ + − ~ ω2(ˆa +a ˆ†)4, (8) 0.05 T=1K ~ 2 4

Differential Capacitance ( fF / μ m T=0K where 0.01 π3S 5v6 χ4 π 3v2 τ = ~ F = ~ F , (9) 0.05 0.10 0.50 1 5 10 50 4 5 2 3 2 ln (16)(kBT ) 8 ln (16)S(kBT ) V(mV) is a characteristic time constant of nonlinear interaction. FIG. 2. The differential capacitance of the proposed layered Clearly, (8) represents an anharmonic oscillator [15–17] structure at various temperatures on the logarithmic scale. and admits a Kerr-like nonlinearity, and this compares Dashed line is the zero temperature limit. Dotted lines cor- well to that of the Josephson junctions [10–14]. It is respond to the quantum capacitance CQ only, without taking not difficult to verify that the range of validity of the the effect of geometric capacitance CG into account. anharmonic approximation (9) is limited to ~ω ≈ e|V | < 2kBT , which by expressing T in K and frequency in GHz gives the upper photon number limit of nmax = 41.7T/f. It is remarkable that the quantum capacitance (5) is Hence, at T = 1K and frequencies in the GHz range, inherently independent of the dielectric thickness. On many photons may interact anharmonically. Beyond this one hand, a dielectric too thin would allow tunneling value, the 6th order nonlinearity kicks in, which remains currents which appear as a parallel nonlinear resistance valid in the range 4kBT < e|V | < 6kBT , and so on for across the capacitor. This might be useful in some ap- higher order interactions [47]. Similarly, at sub-Kelvin plications such as light emission [45, 46], however it is temperatures, the usefulness of this nonlinear capacitor highly non-desirable for this type of circuit element and may become limited to sub-GHz photons. It should be must be avoided. Hence, a minimum dielectric thick- stressed out that one could access a 3-rd order interaction ness of roughly 3nm is needed to ensure that tunneling according to (3) via appropriate DC voltage bias. currents are fully blocked. On the other hand, a dielec- It is clear from the above that the nonlinearity is pro- tric too thick would cause quantum capacitance to fade nounced at small capacitor areas and low temperature. away while appearing in series with the geometric capac- We shall observe that at the typical GHz microwave fre- itance. The reason is that it is the geometric capacitance quencies and temperatures below 1K this expression eas- which decreases significantly at small dielectric thickness. ily gives rise to a quite significant Kerr interaction. Quantum regime is accessible only if the quantum capac- Next, we may define a measure of anharmonicity for itance is much smaller than the geometric capacitance. applications in qubit design [48] as A = |1 − (ω21/ω10)| × These criteria set the above mentioned boundaries on the 100%, where ωmn is the transition frequency between minimum and maximum thickness of the insulating bar- levels m and n. A straightforward calculation gives the rier. estimate A = 12τω. Putting T in K, f in GHz, and S in µm2 gives A = 42.85%f/ST 3. Using the typical values T = 1K, f = 4GHz, and S = 100µm2 gives an anhar- Quantized Hamiltonian monicity of A = 1.1714%, while at T = 0.5K we obtain A = 13.71%, being sufficiently large for qubit applica- Now, suppose that the capacitor is driven by a sinu- tions. soidal voltage with angular frequency ω and a fixed phase As long as the first 0 → 1 and second 1 → 2 transitions φ. The corresponding charge density operator is given as are sufficiently separated in frequency, one could ensure that only exactly one photon with frequency ω could eψ 01 Qˆ = eNˆ = √ (â +a ˆ†), (6) exist in the system. This is how an ideal qubit should 2 operate [15–17], since only the photon states |0i and |1i could be excited by interaction with a radiation reservoir wherea ˆ is the bosonic field operator, satisfying [â, aˆ†] = at frequency ω01. To achieve this, the anharmonicity A 1, and ψ is the fluctuation amplitude of corresponding should be only much larger than the normalized linewidth single-photon number density. Hence, ψ is obtained from of the radiation ∆ω/ω01. Practically, the latter is less (4) as than a hundredth of a percent, so an anharmonicity of s A = 13.7% is quite large. As a result, creation of super- k T ln(16) √ B position states |αi = c0 |0i + c1 |1i with cj, j = 1, 2 being ψ = 2 ω = χ ω. (7) 2πS~vF complex constants and hα|αi = 1, becomes possible. 4

By setting a parallel linear resonant inductor and an remaining significant interactions. The RWA approx- optional series linear capacitor, this lays down the foun- imation is not valid in general, and in some cases, dation of a novel type of qubit based on nonlinear quan- the counter-rotating terms have been shown to have a tum capacitance without need to external magnetic bias. considerable effect on the evolution of system [49, 50]. This is to be compared with other types of nonlinear Such a Hamiltonian has been achieved in systems us- inductances, such as Josephson junctions [10–14] and ki- ing Josephson junction [63] and extended to realise non- netic inductance [8, 9]. Since, a GHz microwave qubit reciprocal devices [62]. Anyhow, by application of RWA should typically operate at cryogenic temperatures, the one may distinguish two particular resonant cases with area of the quantum capacitance should probably be in- 2Ω = ω1 ±ω2, for which the interaction Hamiltonian (11) creased to avoid overly strong interaction. It remains, greatly simplifies and takes either of the forms however, a matter of speculation that whether this pro- 2jθ † −2jθ † posed type of qubits could sustain their coherence over H ≈ ~G(e aˆ1aˆ2 + e aˆ1aˆ2), (14) longer time intervals comparing to other superconducting for 2Ω = ω1 − ω2 and technologies. 2jθ † † −2jθ H ≈ ~G(e aˆ1aˆ2 + e aˆ1aˆ2), (15) for 2Ω = ω + ω . Also, G = g |a¯|2 = 3γ |a¯|2 is the Multi-mode Fields 1 2 0 012 overall interaction rate and θ = ∠a¯ is the phase of the coherent drive. In a similar manner, for a multi-mode field composed of On one hand, by choice of 2Ω = ω1 −ω2 and the Hamil- a few sinusoidal terms and by adding resonant√ inductive tonian (14), one may achieve a beam-splitter or hopping 1 ˙2 ˆ terms as 2 ΣLnqˆn withq ˆn = SQn and ωn = 1/ SLnC0, interaction, where the mode at frequency Ω is pumped we may write down the total Hamiltonian strongly by a coherent field. This type of interaction can facilitate, for example, a non-reciprocal response between " #4 ˆ X † 1 ~τ X √ † the modes under the appropriate conditions [60, 64]. H = ~ ωn aˆnaˆn + − ωn(ân +a ˆn) . 2 4 On the other hand, by choice of 2Ω = ω1 + ω2 and the n n Hamiltonian (15), one may achieve the parametric ampli- (10) fier interaction. We can set up a system, similar to that Here, we have the usual commutation rules between indi- discussed in [62], in order to realise a directional ampli- vidual field operators as [â , aˆ† ] = δ and [â , aˆ ] = 0. n m nm n m fier. However, the coupling is mediated by the quantum Now, we limit the discussion to only three modes, where capacitor, as opposed to a Josephson junction, which in the interacting and non-interacting Hˆ terms may be H 0 principle requires a magnetic bias to induce the third or- separated as Hˆ = Hˆ − to yield the three basic types 0 H der interaction in the system. of interactions in the Hamiltonian Remarkably, this strategy to select the desirable type (1) (2) (3) of interaction by appropriate selection of pump frequency H = H + H + H , (11) has been verified experimentally in a recent experiment with [65], where non-reciprocal transmission of microwave signals is achieved in a nonlinear superconducting circuit (1) 1 X † 2 † 2 = γ (â +a ˆ ) (â +a ˆ ) , (12) based on Nb/Al2O3/Nb Josephson junctions with a bi- H 2 ~ nmm n n m m n6=m ased external DC magnetic field. Thereby, a Field Programmable Josephson Amplifier (FPJA) is obtained (2) 3 X † 3 † = γnmm(ân +a ˆ ) (âm +a ˆ ), which is capable of performing different functionalities. H 2 ~ n m n6=m The clear possible advantage of the proposed nonlinear 3 X capacitor is no need to external magnetic field. (3) = γ (â +a ˆ† )2(â +a ˆ† )(â +a ˆ†). H 2 ~ nml n n m m l l n6=m6=l √ Single-Photon Interaction Here, γnml = τωn ωmωl. We can suppose that the interaction is driven at mode It is possible to estimate the interaction from the above 0 with a strong coherent field and frequency Ω = ω0, relations for G. Expressing the temperature T in K, ca- in such a way thata ˆ0 → a¯ + δaˆ. Neglecting beyond 2 simple nonlinear interactions between modes 1 and 2, pacitor area S in µm , and drive frequency f = Ω/2π, (3) electromagnetic frequencies f = ω /2π and also the and with three modes interacting, only H contributes n n significantly. Then, we obtain single-photon interaction rate g0 in GHz, we obtain √ f f1f2 = G¯(â +a ˆ†)(â +a ˆ†), (13) g = 2π × 0.143 . (16) H ~ 1 1 2 2 0 ST 3 ∗ 2 2 where G¯ = 3γ012(¯a +a ¯ ) . As an example, taking S = 100µm at T = 1K, with Now we use the Rotating Wave Approximation (RWA) f = 4GHz, f1 = 2GHz, and f2 = 10GHz satisfies the res- and neglect the counter-rotating terms, to obtain the onance condition for the hopping interaction (14). This 5 gives a linear capacitance value of C0 = 5.63fF and also current is supplied in the Ohmic regime. Not only for makes it possible to calculate the values of corresponding the structure under consideration, it approaches zero at tank inductors and external circuitry. For these values, zero bias, but also the electric current is mainly non- the single-photon interaction rate is g0 = 2π×25.55MHz. Ohmic and dissipation-free over distances of the order This fairly well falls into the domain of strong coupling. of few microns, since charges move almost ballistically The same value at the temperature of T = 4K is severely across graphene electrodes. So, it should be safe to ig- degraded to the weakly coupled value g0 = 2π×399.2kHz, nore both of these aspects in a precisely fabricated 2D while at the lower temperature of T = 0.25K, it is sig- layered structure, as long as it is void of defects and in- nificantly enhanced to g0 = 2π × 1.635GHz, which is terfaces are atomically flat. It must be pointed out that of course ultrastrongly coupled. Hence, it is possible to these restrictions on fabrication can be experimentally obtain extremely strong single-photon nonlinear interac- challenging to achieve, and are studied thoroughly in a tion rates even with low pump levels, without need to recent article to which the readers are being referred [67]. dilution refrigerators. At lower temperatures, system en- Also, the minimal quantum conductivity σQ may ideally ters the ultrastrongly-coupled regime and g0 can even not be achievable due to the disorder effects and finite exceed the system resonant frequencies [66], while at a temperature [44]. higher temperature above a few Kelvins, the nonlinear- Meanwhile, we note that for the quantum capacitor ity becomes too weak. This is a result of either (9) or under consideration, the electrical current spreads hori- (16), where there is a ∝ S−1T −3 dependence of the an- zontally over the graphene layers. Now since it is the area harmonicity on area and temperature. Furthermore, the S which is dominant, one may proportionally shrink the area S could be freely chosen according to the operation graphene length and increase the electrode width until frequency fj, j = 1, 2 and temperature T to yield the the series resistance is negligible. If the length is suffi- desirable interaction strength g0. Typically for GHz mi- ciently small, being less than mean free path of carriers crowave circuits at liquid Helium temperatures, it could at the cryogenic temperatures, then the expectation that be in the range of 1 − 100µm2. the ballistic propagation of carriers would prevent Ohmic Other issues to be considered here is the superconduc- losses to appear, fully makes sense. It should be added tor contact to the graphene electrodes, as well as the that with the recent discovery of induced superconductiv- charging time of the capacitor. The first issue brings up ity in single-layer graphene [68] with a transition temper- the proximity effect, while the second is related to the ature of 4.2K, hopes are on rise for a fully dissipation-free intrinsic conductivity of the graphene. The proximity operation of the proposed system. effect has been well studied [51–54]. It has been experimentally and theoretically established that in graphene, the finite normal state conductance can maintain finite II. MICROWAVE CIRCULATOR supercurrent, even at zero bias. Hence, the electrical current should have no problem flowing freely through graphene electrodes. Moreover, it is well known that car- Now, as potential applications, we discuss schemes for riers in graphene are essentially ballistic [55, 56] at low two non-reciprocal devices. Assume that initially we have temperatures over distances of a few microns. three electromagnetic modes with frequencies ω1, ω2, and ω3, mutually coupled via three nonlinear 2D capacitors, interacting through the Hamiltonian described in (14). Charging Delay We set up the hopping interaction between each of the 3 modes using the scheme mentioned previously. Hence Next, we may turn to the issue of charging delay. This the Hamiltonian of the system could be written as H = issue has been studied in mesoscopic capacitors [57] and H12 + H23 + H31, where a universal charge resistance has been found for normal iφ3 † −iφ3 † conductors. Remarkably, at zero bias, the steady state H12 = ~g3(e aˆ1aˆ2 + e aˆ1aˆ2), (18) iφ1 † −iφ1 † charge density of graphene is also expected to be zero, H23 = ~g1(e aˆ2aˆ3 + e aˆ2aˆ3), which should admit no normal state conductance, but it † † = g (eiφ2 aˆ aˆ + e−iφ2 aˆ aˆ ). is well known also that the conductance of graphene at H31 ~ 2 3 1 3 1 zero bias is also nonzero and finite, being in fact equal to 2 It is worth mentioning here that the resonance condition the quantum conductance σQ = 2e /π~ [54, 58, 59]. for hopping interaction Hamiltonian, requires the con- In that sense, the ratio of quantum capacitance to straint Ω + Ω = Ω on pump drives. quantum conductance with the unit of time, is simply 1 3 2 The corresponding circuit could be easily set up by proportional to the Fermi energy making a topological dual equivalent circuit [69] of

CQ |EF | the Field-Programmable Josephson Array (FPJA) pro- τQ = S = S 2 . (17) posed recently elsewhere [65] for a superconducting non- σQ ~v F dissipative reciprocal ampliﬁer. This will lead to a cir- The time-constant τQ determines the quantum charac- cuit as shown in Fig. 3, being here referred to as the teristic time for charge/discharge of capacitor when the Field-Programmable Capacitor Array (FPCA), held at 6

Signals + Signal Pumps Output Circulator LNA 3

퐿 1 퐿2 퐿3 푄 푄 퐶1 퐶2 퐶1 퐶2 퐶3

-ϕ2 ϕ1 FIG. 3. Proposed nonlinear circuit for the Field- ℍ ℍ Programmable Capacitor Array. 31 23 cryogenic temperature. There are two nonlinear quantum capacitors CQj, j = 1, 2, joined with six linear superconducting elements, composed of three linear tank inductors Lj, j = 1, 2, 3 and three linear tank capacitors C , j = 1, 2, 3. These determine the resonant frequen- j 1 ℍ 2 cies of the system. Pumps feed through the quantum 12 capacitors at desirable and fixed frequencies to achieve the needed interaction types and strengths. Collected signals are fed through a regular microwave circulator at ϕ3 a higher temperature to a Low Noise Amplifer (LNA), prior to detection and signal analysis. A similar strat- FIG. 4. The mode coupling scheme in an ideal circulator. The egy for a four-mode Diamond-shaped system could be signal can take two pathways for conversion from, say, mode implemented to obtain a non-dissipative microwave su- 1 to mode 3. The phase difference between the two paths, π perconducting circuit, in order to achieve extremely high ∆φ, can be controlled and by setting ∆φ = ± 2 provides the non-reciprocal amplification and isolation, while main- condition for interference between the paths. Hence only one pathway is preferred (say, anti-clockwise, blue curve) and not taining identical input/output frequencies [70]. the opposite(clockwise) for mode conversion. In this scheme, A probe signal (input) at, say mode 1, has two path- any one of the 3 modes can be taken as the input while any ways for conversion to mode 3 (1 → 2 : 2 → 3 or 1 → 3). of the other two may play the role of output. The signal picks up different phases while traversing through these two paths [61]. The phase difference at mode 3 (output) is ∆φ = φ1 +φ3 −φ2 (Fig. 4). Here, the the frequency range of interest. For the simulation done, horizontal axis is the normalized detuning ∆ with respect we have taken ω1 = 2π × 1GHz, ω2 = 2π × 1.05GHz, to the resonant frequency ω. By controlling the phases ω3 = ω1 + ω2, g1 = g2 = g3 = 2π × 1GHz, and of the pumps, one can set ∆φ = ± π , thereby allowing 2 κ1 = κ2 = κ3 = 2g1. interference between the two paths. This results in the Physically, in the two schemes discussed above, ∆φ signal propagating only in one of the two paths, hence acts as an artificial gauge flux and plays the role of the breaking the symmetry of the system. Hence we realise magnetic field in a conventional circulator [62]. The lack an ideal circulator. Using the quantum Langevin equa- of an external magnetic bias is advantageous in situations tion in the RWA and input-output formalism [2, 71, 72], such as qubit read-out schemes, where the performance is one can write the equations of motion, and hence the severely hindered by the unwanted noises that enter the Langevin matrix of the system as system through various ports. Hence such a scheme could

 κ1 −iφ3 iφ2  be a precursor to devices which achieve non-reciprocity − iω1 + 2 −ig3e −ig2e iφ3 κ2 −iφ1 between such physical ports of the system. M =  −ig3e − iω2 + 2 −ig1e  , (19) −iφ2 iφ1 κ3 −ig3e −ig2e − iω3 + 2 CONCLUSIONS where κi is the decay rate of the i-th mode. We see that M is not symmetric about the diagonal. In order to see the non-reciprocal response, one can solve the above sys- In summary, we have shown that a layered 2D quantum tem in the Fourier space, obtaining the scattering matrix capacitor with graphene electrodes is highly nonlinear, if S that relates the input signal to the mode converted out- temperature is held sufficiently low and there is no bias. put signal asa ôut = S aîn. By tuning the parameters, it All requirements were shown to be met at microwave is clear from the ratio of scattering amplitudes that sym- (GHz) frequencies, temperatures of the order of 1K, and metry of the system has indeed been broken, as shown in capacitors as small as a few µm2. We showed that under Fig. 5. Finally, the insertion loss of the system defined as the correct choice of strong pump and weak fields, the 2 IL = −10 log10 |S13| has been also computed and plot- interaction Hamiltonian would be that of a beam split- ted in Fig. 6, showing that it is essentially negligible over ter with adjustable phase from the pump, depending on 7

S 2 13 easily forced into the ultrastrong regime. It is antici- S 2 31 pated that this proposal would open up a new highway 1.0 in the ﬁeld of quantum microwave electro-optics, while allowing access to interaction rates which are not easily 0.8 achievable though other methods. The high strength of Kerr nonlinearity makes the proposed quantum capacitor 0.6 potentially very useful for design of novel charge quantum bits as well. 0.4

0.2 ACKNOWLEDGMENT Δ -10 -5 5 10 ω The authors wish to thank Dr. Alexey Feofanov and Dr. Dalziel Wilson at Ecole´ Polytechnique Fédéralede FIG. 5. The ratio of between 1 → 3 and 3 → 1 mode conver- Lausanne, and Dr. Simon Nigg at the University of Basel sion scattering amplitudes. Due to destructive interference, for reading the manuscript and making useful sugges- set up by the phase difference parameter, the symmetry of the system is broken. tions. CONTRIBUTIONS

Insertion Loss(dB) The idea and analysis of 2D quantum capacitor was

2.5 conceived by S. K. Circulator design and analysis was done by A. K. Both authors discussed the results and contributed signiﬁcantly to the write up of the paper. 2.0

1.5 COMPETING INTERESTS

The authors declare no competing ﬁnancial interests. 1.0

Δ -1.0 -0.5 0.5 0.5 1.0 ω FUNDING

FIG. 6. Illustration of insertion loss for the proposed circulator. S. K. is being supported in part by Research Deputy of Sharif University of Technology over a sabbatical visit. This research has been supported in part by the Labora- the resonance condition. Potential applications were dis- tory of Photonic and Quantum Measurements at Ecole´ cussed and it was shown that the interaction could be Polytechnique F´ed´eralede Lausanne.

[1] Kippenberg, T. J. & Vahala, K. J. Cavity optomechan- arxiv: 1602.05180 (2016). ics: back-action at the mesoscale. Science 321, 1172-1176 [8] Salahuddin, S. & Lundstrom, M. Transport effects on sig- (2008). nal propagation in quantum wires. IEEE Trans. Electron [2] Aspelmeyer, M, Kippenberg, T. & Marquardt, F. Cavity Dev. 52, 1734-1742 (2005). Optomechanics, (Springer, Berlin, 2014). [9] Annunziata, A. J., Santavicca, D. F., Frunzio, L., Cate- [3] Aspelmeyer, M., Kippenberg, T. J., & Marquardt, F. lani, G., Rooks, M. J., Frydman, A. & Prober, D. E. Cavity optomechanics. Rev. Mod. Phys. 86, 1391 (2014). Tunable superconducting nanoinductors. Nanotechnology [4] Marquardt, F., & Girvin, S. M. Optomechanics. Physics 21, 445202 (2010). 2, 40 (2009). [10] Sillanpää, M. A., Lehtinen, T., Paila, A., Makhlin, [5] Meystre, P. A short walk through quantum optomechan- Yu., Roschier, L. & Hakonen, P. J. Direct Observation ics. Ann. Phys. 525, 215-233 (2013). of Josephson Capacitance. Phys. Rev. Lett. 95, 206806 [6] Khorasani, S. Coupled Mode Theory of Optomechani- (2005). cal Crystals. IEEE J. Quantum Electron. 52, 6100406 [11] Pirkkalainen, J.-M., Cho, S. U., Massel, F., Tuorila, J., (2016). Heikkil, T. T., Hakonen, P. J. & Sillanpää,M. A. Cavity [7] Tóth,L. D., Bernier, N. R., Nunnenkamp, A., Feofanov, optomechanics mediated by a quantum two-level system. A. K. & Kippenberg, T. J. A dissipative quantum reser- Nat. Commun. 6, 6981 (2015). voir for microwave light using a mechanical oscillator. 8

[12] Barends, R., Shabani, A., Lamata, L., Kelly, J., Mez- [28] Tsang, M. Cavity quantum electro-optics. II. Input- zacapo, A., Las Heras, U., Babbush, R., Fowler, A. output relations between traveling optical and microwave G., Campbell, B., Chen, Yu, Chen, Z., Chiaro, B., fields. Phys. Rev. A 84, 043845 (2011). Dunsworth, A., Jeffrey, E., Lucero, E., Megrant, A., Mu- [29] Rozhkov, A. V., Giavaras, G., Bliokh, Y. P., Frei- tus, J. Y., Neeley, M., Neill, C., O’Malley, P. J. J., Quin- likher, V. & Nori, F. Electronic properties of mesoscopic tana, C., Roushan, P., Sank, D., Vainsencher, A., Wen- graphene structures: Charge confinement and control of ner, J., White, T. C., Solano, E., Neven, H. & Martinis, spin and charge. Phys. Rep. 503, 77-114 (2011). J. M. Digitized adiabatic quantum computing with a su- [30] Rozhkov, A. V., Savelev, S. & Nori, F. Electronic prop- perconducting circuit. Nature 534, 222-226 (2016). erties of armchair graphene nanoribbons. Phys. Rev. B [13] Oboznov, V. A., Bo´lginov, V. V., Feofanov, A. K., 79, 125420 (2009). Ryazanov, V. V. & Buzdin, A. I. Thickness Dependence [31] Ponomarenko, L. A., Yang, R., Gorbachev, R. V., Blake, of the Josephson Ground States of Superconductor- P., Mayorov, A. S., Novoselov, K. S., Katsnelson, M. Ferromagnet-Superconductor Junctions. Phys. Rev. Lett. I. & Geim, A. K. Density of States and Zero Landau 96, 197003 (2006). Level Probed through Capacitance of Graphene. Phys. [14] Feofanov, A. K., Oboznov, V. A., Bo´lginov, V. V., Lisen- Rev. Lett. 105, 136801 (2010). feld, J., Poletto, S., Ryazanov, V. V., Rossolenko, A. [32] Yu, G. L., Jalil, R., Belle, B., Mayorov, A. S., Blake, P., N., Khabipov, M., Balashov, D., Zorin, A. B., Dmitriev, Schedin, F., Morozov, S. V., Ponomarenko, L. A., Chi- P. N., Koshelets, V. P. & Ustinov, A. V. Implementa- appini, F., Wiedmann, S., Zeitler, U. Katsnelson, M. I., tion of superconductor/ferromagnet/superconductor π- Geim, A. K., Novoselov, K. S. & Elias, D. C. Interaction shifters in superconducting digital and quantum circuits. phenomena in graphene seen through quantum capaci- Nat. Phys. 6, 593-597 (2010). tance. Proc. Nat. Acad. Sci. 110, 3282-3286 (2013). [15] You, J. Q. & Nori, F. Atomic physics and quantum op- [33] Deen, D. A., Olson, E. J., Ebrish, M. A. & Koester, S. J. tics using superconducting circuits. Nature 474, 589-597 Graphene-Based Quantum Capacitance Wireless Vapor (2011). Sensors. IEEE Sensors J. 14, 1459-1466 (2013). [16] Georgescu, I. M., Ashhab, S. & Nori, F. Quantum simu- [34] Koester, S. J. High quality factor graphene varactors lation. Rev. Mod. Phys. 86, 153 (2014). for wireless sensing applications. Appl. Phys. Lett. 99, [17] Xiang, Z.-L., Ashhab, S., You, J. Q. & Nori, F. Hybrid 163105 (2016). quantum circuits: Superconducting circuits interacting [35] Duty, T., Johansson, G., Bladh, K., Gunnarsson, D., with other quantum systems. Rev. Mod. Phys. 85, 623 Wilson, C. & Delsing, P. Observation of Quantum Ca- (2013). pacitance in the Cooper-Pair Transistor. Phys. Rev. Lett. [18] Kim, E.-J., Johansson, J. R. & Nori, F. Circuit analog 95, 206807 (2005). of quadratic optomechanics. Phys. Rev. A 91, 033835 [36] Sillanpää, M. A., Lehtinen, T., Paila, A., Makhlin, (2015). Y., Roschier, L. & Hakonen, P. J. Direct Observation [19] Shevchenko, S. N., Pershin, Y. V. & Nori, F. Qubit-Based of Josephson Capacitance. Phys. Rev. Lett. 95, 206806 Memcapacitors and Meminductors. Phys. Rev. Appl. 6, (2005). 014006 (2016). [37] Shevchenko, S. N., Ashhab, S. & Nori, F. Inverse Landau- [20] Likharev, K. K. Dynamics of Josephson Junctions and Zener-Stckelberg problem for qubit-resonator systems. Circuits (Gordon and Breach, Amsterdam, 1983). Phys. Rev. B 85, 094502 (2012). [21] Song, X., Oksanen, M., Li, J., Hakonen, P. J. & Sil- [38] Shevchenko, S. N., Rubanov, D. G. & Nori, F. Delayed- lanpää, M. A. Graphene Optomechanics Realized at response quantum back action in nanoelectromechanical Microwave Frequencies. Phys. Rev. Lett. 113, 027404 systems. Phys. Rev. B 91, 165422 (2015). (2014). [39] Nigg, S. E., Loerch, N. & Tiwari, R. P. Robust quan- [22] Singh, V., Shevchuk, O., Blanter, Ya. M. & Steele, G. tum optimizer with full connectivity. arxiv: 1609.06282 A. Negative nonlinear damping of a multilayer graphene (2016). mechanical resonator. Phys. Rev. B 93, 245407 (2016). [40] Dröscher, S., Roulleau, P., Molitor, F., Studerus, P., [23] Cole, R. M., Brawley, G. A., Adiga, V. P., De Alba, R., Stampfer, C., Ensslin, K. & Ihn, T. Quantum capaci- Parpia, J. M., Ilic, B., Craighead, H. G. & Bowen, W. P. tance and density of states of graphene. Appl. Phys. Lett. Evanescent-Field Optical Readout of Graphene Mechan- 96, 152104 (2010). ical Motion at Room Temperature. Phys. Rev. Applied [41] Fang, T., Konar, A., Xing, H. & Jena, D. Carrier statis- 3, 024004 (2015). tics and quantum capacitance of graphene sheets and rib- [24] Khorasani, S. Tunable spontaneous emission from layered bons. Appl. Phys. Lett. 91, 092109 (2007). graphene/dielectric tunnel junctions. IEEE J. Quantum [42] Xia, J., Chen, F., Li, J. & Tao, N. Measurement of the Electron. 50, 307-313 (2014). quantum capacitance of graphene. Nat. Nanotech. 4, 505- [25] Palacios-Berraquero, C., Barbone, M., Kara, D. M., 509 (2009). Chen, X., Goykhman, I., Yoon, D., Ott, A. K., Beit- [43] Martin, J., Akerman, N., Ulbricht, G., Lohmann, T., ner, J., Watanabe, K., Taniguchi, T., Ferrari, A. C. Smet, J. H., von Klitzing, K., & Yacoby, A. Observa- & Atatüre,M. Atomically thin quantum light-emitting tion of electronhole puddles in graphene using a scan- diodes. Nat. Commun. 7, 12978 (2016). ning single-electron transistor. Nature Phys. 4, 144-148 [26] Caprara Vivoli, V., Barnea, T., Galland, C. & San- (2008). gouard, N. Proposal for an Optomechanical Bell Test. [44] Kotov, V. N., Uchoa, B., Pereira, V. M., Guinea, F. Phys. Rev. Lett. 116, 070405 (2016). & Castro Neto, A. H. Electron-Electron Interactions in [27] Tsang, M. Cavity quantum electro-optics. Phys. Rev. A Graphene: Current Status and Perspectives. Rev. Mod. 81, 063837 (2010). Phys. 84, 1067 (2012). 9

[45] Phare, C. T., Daniel Lee, Y.-H., Cardenas, J. & Lipson, ing. Phys. Rev. X 5, 021025 (2015). M. Graphene electro-optic modulator with 30GHz band- [61] Ranzani, L. & Aumentado, J. Graph-based analysis of width. Nat. Photon. 9, 511-514 (2015). nonreciprocity in coupled-mode systems. New J. Phys. [46] Sun, Z., Martinez, A. & Wang, F. Optical modulators 17, 023024 (2015). with 2D layered materials. Nat. Photon. 10, 227-238 [62] Sliwa, K. M., Hatridge, M., Narla, A., Shankar, S., Frun- (2016). zio, L., Schoelkopf, R. J. & Devoret, M. H. Reconfig- [47] Khorasani, S. Higher order interactions in quantum op- urable Josephson Circulator/Directional Amplifier. Phys. tomechanics. arxiv: 1610.08780 (2016). Rev. X 5, 041020 (2015). [48] Clarke, J. & Wilhelm, F. K. Superconducting quantum [63] Abdo, B., Kamal, A. & Devoret, M. H. Nondegenerate bits. Nature 453, 1031-1042 (2008). three-wave mixing with the Josephson ring modulator. [49] Alidoosty, M., Khorasani, S. & Aram, M. H. Simulation Phys. Rev. B 87, 014508 (2013). of Multipartite Cavity Quantum Electrodynamics. IEEE [64] Nunnenkamp, A., Sudhir, V., Feofanov, A. K., Roulet, J. Quantum Electron. 49, 1066-1079 (2013). A. & Kippenberg, T. J. Quantum-Limited Amplification [50] Malz, D. & Nunnenkamp, A. Optomechanical dual-beam and Parametric Instability in the Reversed Dissipation backaction-evading measurement beyond the rotating- Regime of Cavity Optomechanics. Phys. Rev. Lett. 113, wave approximation. Phys. Rev. A 94, 053820 (2016). 023604 (2014). [51] Heersche, H. B., Jarillo-Herrero, P., Oostinga, J. B., Van- [65] Lecocq, A., Ranzani, L., Peterson, G. A., Cicak, K., dersypen, L. M. K. & Morpurgo, A. F. Bipolar supercur- Simmonds, R. W., Teufel, J. D. & Aumentado, J. Non- rent in graphene. Nature 446, 56-59 (2007). reciprocal microwave signal processing with a Field- [52] Ojeda-Aristizabal, C., Ferrier, M., Guéron,S. & Bouch- Programmable Josephson Amplifier. arxiv: 1612.01438 iat, H. Tuning the proximity effect in a superconductor- (2016). graphene-superconductor junction. Phys. Rev. B 79, [66] Yoshihara, F., Fuse, T., Ashhab, S., Kakuyanagi, K., 165436 (2009). Saito, S. & Semba, K. Superconducting qubitoscillator [53] Komatsu, K., Li, C., Autier-Laurent, S., Bouchiat, H. & circuit beyond the ultrastrong-coupling regime. Nature Guéron,S., “Superconducting proximity effect in long su- Phys. 13, 44-47 (2017). perconductor/graphene/superconductor junctions: From [67] Moldovan, C. F., Vitale, W. A., Sharma, P., Tamagnone, specular Andreev reflection at zero field to the quantum M., Mosig, J. R. & Ionescu, A. M. Graphene Quantum Hall regime. Phys. Rev. B 86, 115412 (2012). Capacitors for High Frequency Tunable Analog Applica- [54] Li, C., Guéron, S., Chepelianskii, A. & Bouchiat, H. tions. Nano Lett. 16, 4746-4753 (2016). Full range of proximity effect probed with supercon- [68] Di Bernardo, A., Millo, O., Barbone, M., Alpern, H., ductor/graphene/superconductor junctions. Phys. Rev. Kalcheim, Y., Sassi, U., Ott, A. K., De Fazio, D., Yoon, B 94, 115405 (2016). D., Amado, M., Ferrari, A. C., Linder, J. & Robinson, [55] Bliokh, Y. P., Freilikher, V. & Nori, F. Ballistic charge J. W. A. p-wave triggered superconductivity in single- transport in graphene and light propagation in periodic layer graphene on an electron-doped oxide superconduc- dielectric structures with metamaterials: A comparative tor. Nature Commun. 8, 14024 (2017). study. Phys. Rev. B 87, 245134 (2013). [69] Desoer, C. A. & Kuh, E. S. Basic Circuit Theory [56] Rozhkov, A. V., Sboychakov, A. O., Rakhmanov, A. L. & (McGraw-Hill, New York, 1969). Nori, F. Electronic properties of graphene-based bilayer [70] Khorasani, S. Diamond Configuration for Non-reciprocal systems. Phys. Rep. 648, 1-104 (2016). Transmission. arxiv: 1612.01836 (2016). [57] Nigg, S. E., López, R. & Büttiker, M. Mesoscopic Charge [71] Gardiner, C. W. & Collett, M. J. Input and output in Relaxation. Phys. Rev. Lett. 97, 206804 (2006). damped quantum systems: Quantum stochastic differen- [58] Li, T. C. & Lu, S.-P. Quantum conductance of graphene tial equations and the master equation. Phys. Rev. A 31, nanoribbons with edge defects. Phys. Rev. B 77, 085408 3761-3774 (1985). (2008). [72] Clerk, A. A., Devoret, M. H., Girvin, S. M., Marquardt, [59] Tombros, N., Veligura, A., Junesch, J., Guimarães,M. F. & Schoelkopf, R. J. Introduction to quantum noise, H. D, Vera-Marun, I. J., Jonkman, H. T. & van Wees, measurement, and amplification. Rev. Mod. Phys. 82, B. J. Quantized conductance of a suspended graphene 1155 (2010). nanoconstriction. Nat. Phys. 7, 697-700 (2011). [60] Metelmann, A. & Clerk, A. A. Nonreciprocal Photon Transmission and Amplification via Reservoir Engineer-