Arxiv:2008.08268V2 [Quant-Ph] 4 Aug 2021

Photon-number-dependent eﬀective Lamb shift

Arto Viitanen,1, ∗ Matti Silveri,2 MátéJenei,1, 3 Vasilii Sevriuk,1, 3 Kuan Y. Tan,1, 3 Matti Partanen,1 Jan Goetz,1, 3 Leif Grönberg,4 Vasilii Vadimov,1, 5, 6 Valtteri Lahtinen,1 and Mikko Möttönen1, 4, ∗ 1QCD Labs, QTF Center of Excellence, Department of Applied Physics, Aalto University, P.O. Box 13500, FI-00076 Aalto, Finland. 2Research Unit of Nano and Molecular Systems, University of Oulu, P.O. Box 3000, FI-90014 Oulu, Finland. 3IQM Finland Oy, Keilaranta 19, FI-02150 Espoo, Finland. 4VTT Technical Research Centre of Finland Ltd., QTF Center of Excellence, P.O. Box 1000, FI-02044 VTT, Finland. 5MSP Group, QTF Centre of Excellence, Department of Applied Physics, Aalto University, P.O. Box 11000, FI-00076 Aalto, Finland 6Institute for Physics of Microstructures, Russian Academy of Sciences, 603950 Nizhny Novgorod, GSP-105, Russia (Dated: August 5, 2021) The Lamb shift, an energy shift arising from the presence of the electromagnetic vacuum, has been observed in various quantum systems and established as the part of the energy shift independent of the environmental photon number. However, typical studies are based on simplistic bosonic models which may be challenged in practical quantum devices. We demonstrate a hybrid bosonic-fermionic environment for a linear resonator mode and observe that the photon number in the environment can dramatically increase both the dissipation and the effective Lamb shift of the mode. Our observations are quantitatively described by a first-principles model which we develop here also to guide device design for future quantum-technological applications. The device demonstrated here can be utilized as a fully rf-operated quantum-circuit refrigerator to quickly reset superconducting qubits.

I. INTRODUCTION In fact, the Lamb shift has not only been studied in natural quantum systems such as atoms [10, 11, 15–18], Quantum theory, put forward a century ago has not only but has also fascinated physicists working on engineered revealed the deepest secrets of nature such as those of the quantum systems. The Lamb shift in superconducting elementary particles but also given rise to a multitude of qubits was first observed as a result of a single bosonic practical applications ranging from medical imaging [1, 2] mode [19, 20] and later arising from a few well-defined to the emerging quantum computers [3–7] and quantum modes [21]. Also the Lamb shift owing to lattice vibra- cryptography [8, 9]. One of the most important challenges tions, or phonons, has been observed in ultracold quan- in contemporary quantum physics is the understanding tum gases [18]. Recently, the studies of the Lamb shift in and control of the effects of environments on the studied microwave circuits were taken a step closer to Lamb’s orig- quantum systems. inal experiments, namely, to broadband environments [22]. In 1947, Willis Lamb and Robert Retherford conducted This was achieved with a quantum-circuit refrigerator [23] their celebrated experiments to observe an anomalous (QCR) which was used as a tunable environment consist- small shift in the energy spectrum of hydrogen [10]. This ing of a continuum of modes. frequency shift, later referred to as the Lamb shift, was The Lamb shift originates from the coupling between considered to arise from the sole presence of the electro- the environmental modes and the studied system, even magnetic vacuum, not from excitations of the field [11]. though the modes are not excited [10, 11, 24]. In contrast, It paved the way for the golden ages of physics where the shift that stems from the occupation of the modes modern quantum electrodynamics and particle physics is generally referred to as the ac Stark shift [24]. If a arXiv:2008.08268v2 [quant-ph] 4 Aug 2021 were developed along with discoveries such as a theory transition frequency of the studied system depends non- of superconductivity [12], the Casimir effect [13], and linearly on a coupled external field, fluctuations in the applications of quantum tunnelling [14]. Yet, the Lamb field give rise to a different average frequency compared shift still ties the contemporary quantum physicists and with the uncoupled system. Thus the higher the occu- engineers with the importance of the quantum environ- pation in the environmental modes, the more the field it ment: quantum systems are not isolated—even if energy induces effectively fluctuates, and the more ac Stark shift relaxation can be mitigated by filtering at the system is observed. Importantly, the ac Stark shift is absent in frequency, the Lamb shift arising from the full spectrum a linear system linearly coupled to its environment [24]. of the environmental modes persists. Thus such a linear system is ideal for studying the Lamb shift which persists. Typically, as observed in the case of linear bosonic environments [10, 11, 15–18, 22, 25], the Lamb shift depends ∗ To whom correspondence should be addressed; E-mail: on the coupling strength between the system and the en- arto.viitanen@aalto.fi and mikko.mottonen@aalto.fi. vironment and on the density of the environmental states, 2

FIG. 1. Studied system and measurement scheme. (a) Illustration of the interaction between the primary bosonic mode at angular frequency ωp and its engineered environment (dashed box) consisting of a quantum-circuit refrigerator (QCR) and a supporting mode at ωs. (b) Schematics of a coplanar waveguide resonator (green) coupled to the normal-metal island of the QCR on the left and to the measurement setup on the right. The dash-dotted lines show the voltage amplitudes of the primary and supporting modes. A vector network analyzer (VNA) measures the reflection coefficient S11 (see AppendixB). The energy diagram of the QCR illustrates photon-assisted electron tunnelling at a normal-metal–insulator–superconductor (NIS) junction. The electron occupation in the normal-metal (brown shading) follows the Fermi distribution, whereas the superconductor states are essentially filled (blue shading) up to the Bardeen–Cooper–Schrieffer energy gap of 2∆. States at high energies ε are vacant. Photon-assisted tunnelling events can absorb photons from (blue arrows) or emit to (red arrow) the primary mode, of energy ~ωp, or from the supporting mode, of energy ~ωs. Elastic events (black arrow) do not directly induce dissipation on the resonator. Energetically forbidden processes at the example dc bias voltage, denoted by V , are crossed out. but not on the photon number or the temperature of the environment. In this paper, however, we implement a hybrid [26–28] environment with bosonic and fermionic constituents. Namely, we couple our studied system, which is a mode of a superconducting resonator, to the environment consisting of a QCR and another mode of the resonator (Figs.1 and2). The system mode is referred to as the primary mode and the environmental mode as the supporting mode. We measure the resonance frequency FIG. 2. Lumped-element circuit model of the studied system. of the primary mode at different QCR bias voltages and The primary and the supporting mode are each modeled by average photon numbers of the supporting mode (Fig.3). an LC resonator (dark and bright red) with capacitances Cp We observe that the frequency strongly depends on not and Cs, and inductances Lp and Ls, respectively. The res- only the bias voltage [22] but also on the photon number onators are coupled to the normal-metal island of the QCR (Fig.4). This observation of a photon-number-dependent (brown color) through capacitances Cc,p = Cc,s. Tunnelling frequency shift apparently contradicts the typical non- occurs between the normal metal and the superconducting hybrid case highlighting the observed novel physics owing electrode (grey color). This junction is associated with the capacitance C and biased with V , allowing voltage control to the hybrid nature of the environment: an increase j over the photon-assisted electron tunnelling. The effects of the in the photon number in the bosonic part of the envi- other parallel NIS junction (dashed lines) are included in the ronment effectively manifests in the primary mode as capacitance Cm of the normal-metal island to the ground and an increased environmental coupling strength mediated in the value of the junction resistance used in the model. In by the fermionic part, and consequently as an increased the experimental sample, we have a single physical resonator, effective Lamb shift. Finally, we quantitatively explain which implies that in this lumped-element model we have to using first-principles quantum theory (see AppendicesD use equal output capacitances Cg,p = Cg,s and characteris- andF) the origin of a novel oscillatory behavior of the tic impedances Ztr,p = Ztr,s to correctly model the external effective Lamb shift as a function of the bias voltage. coupling of the considered resonator modes. The classical node fluxes at the island and at the resonators are denoted by ΦN, Φp, and Φs, and their conjugate charges by QN,Qp, and Qs, respectively. II. RESULTS

A. Radio frequency quantum-circuit refrigerator junctions with two superconducting leads. In general, a QCR provides tunable dissipation for the electromagnetic Our physical system is described in Figures1 and2 excitations of the coupled quantum circuit which is in (see also Fig.7 in AppendixA) with parameters listed in our case the CPW resonator with two considered modes. TableI. The sample consists of a superconducting coplanar The dissipation is controlled with a bias voltage applied waveguide (CPW) resonator capacitively coupled to a through the two leads equipped with on-chip low-pass QCR, i.e., to a normal-metal island that forms tunnel RC ﬁlters in order to keep the dc line environment well 3

continuously with the tunable bias voltage, yielding eV , TABLE I. Key device and model parameters. See AppendixC n for details of the experimental determination of the parameters. and in a quantized manner by p absorbed photons from the fundamental mode, yielding n ω . With the inclu- Parameter Symbol Value Unit p~ p sion of the rf-driven second mode (ωs/2π = 17.651 GHz, Resonator frequency ωp/2π 8.8241 GHz linewidth 11 MHz), slightly shifted from 2×ωp due to the Supporting mode frequency ωs/2π 17.651 GHz coupling capacitance, the bias voltage is no longer strictly Resonator characteristic impedance Zp 42.8 Ω External coupling strength γtr,p/2π 2.1 MHz necessary to activate tunnelling events even for vanishing Excess coupling strength γ0,p/2π 1.6 MHz np~ωp. Instead, the tunnelling electrons may acquire the Coupling capacitance Cc,p 780 fF necessary energy, in excess of np~ωp, from the photons of Output capacitance Cg,p 6.4 fF the second bosonic mode. Hence, we refer to the funda- Island capacitance CΣ 4 fF mental mode as the primary mode (subscripts p) and to Primary-mode capacitance ratio αp 0.7817 the second mode as the supporting mode (subscripts s). Supporting-mode capacitance ratio αs 0.7413 Superconductor gap parameter ∆ 208 µeV An electron-tunnelling event at the junctions induces −4 Dynes parameter γD 4 × 10 a charge shift of ∆Q = αp/se on the resonator modes. Electron temperature TN 90 mK The strength of the effect is dependent on the circuit parameters (see Fig.2) and determined by the constant 0 < αp/s < 1, which is related to the capacitances of the controlled. The fabrication of the sample is discussed in resonator modes, Cp/s, and of the couplings to the QCR, AppendixA. Cc,p/s (see AppendixD). The charge shift couples the As depicted in Fig.1(b), the interaction between the different eigenstates of the two modes and may induce fundamental mode of the resonator (resonance frequency transitions between them, resulting in photon-assisted ωp/2π = 8.8241 GHz ± 0.18 MHz, linewidth 1.9 MHz) electron tunnelling with annihilation or creation of pho- and its engineered environment is mediated by photon- tons in the two modes, as illustrated in Fig.1(b). assisted electron-tunnelling events through the normal- metal–insulator–superconductor (NIS) junctions in the For the sake of simplicity, we assume that the two NIS QCR. Instead of a single NIS junction, we employ a pair junctions are identical, the corresponding electrodes are of junctions to double the tunnelling rate as well as to at equal temperatures, and the charging energy of the supply a return path to the bias current from the normal- normal-metal island, EN, is small compared to the other metal island. The tunnelling processes are activated if relevant energy scales. After tracing out the supporting the electrons are supplied with sufficient additional en- mode, the electromagnetic environment of the primary ergy to compensate for the superconductor energy gap of mode is characterized by its effective coupling strength 2∆ around the Fermi level. The energy can be provided to the hybrid environment (see AppendixD)

Z 2 2 p X (s) X γT,p(V, n¯s) = 2παp Pk(¯ns) Mkl `pF (τeV + `p~ωp + `s~ωs − EN), (1) RT k,l `p,τ=±1

where Zp is the characteristic impedance of the primary B. Multiphoton-assisted electron tunnelling mode, RT is the tunnelling resistance, Pk is the occupation probability of the kth supporting-mode eigenstate, n¯s = P The argument of the normalized forward tunnelling k kPk is the mean supporting-mode photon number, rate F in equation (1) is the energy obtained in the (s) Mkl is the transition matrix element from the kth to tunnelling process by the electrons from the resonator the lth supporting-mode eigenstate, `s = k − l is the modes and from the voltage source. The remaining en- number of supporting-mode photons absorbed by the ergy for tunnelling is provided by thermal excitations, electron-tunnelling event, F is the normalized forward described by the Fermi functions in the superconducting tunnelling rate (see AppendixD), and τ takes into account electrodes and the normal-metal island (see AppendixD). both direction of the electron tunneling through the NIS Figure3(a) shows three characteristic regions [ 29] for junction. We excite the supporting mode with a classical F (δE): (i) For δE ∆, the electron tunnelling and k −n¯s n¯s γ drive, yielding Pk(¯ns) = e . the resulting coupling strength T,p are suppressed by k! the gap in the density of states of the superconductor. (ii) Once δE is raised suﬃciently near the gap edge, the photon-assisted electron tunnelling is thermally activated, and the coupling strength increases exponentially, (iii) before saturating near the gap edge. The driven supporting mode promotes multiphoton- 4

FIG. 3. Onset of multiphoton processes. (a) Relaxation rate of the primary mode from the state |1i to |0i through processes involving `s = 0, 1, 2, 3 supporting-mode photons as a function of the bias voltage for small (solid lines, ms = 3) and large (dashed lines, ms = 1000) initial supporting-mode occupation ms. Top axis shows the energy needed from the supporting-mode photons to activate the QCR for the primary mode. (b) Reflection magnitude as a function of the bias voltage and probe frequency from negligible supporting-mode drive strength (top left panel) to strong drive (bottom right panel) with the drive power levels at the sample input and the estimated supporting-mode steady-state average photon number n¯s(V = 0) indicated. The reflection magnitude is normalized such that its maximum value is one in each panel. The background has been corrected as described in AppendixC. The arrows indicate the bias points, V = (∆ − ~ωp − `s~ωs)/e, where relaxation processes involving `s = 0 (blue), 1 (red), and 2 (orange) supporting-mode photons become active. assisted electron-tunnelling events, as illustrated in owing to any excess sources, yielding the internal quality Fig.1(b) and quantified in Fig.3(a). Namely, the par- factor of the primary mode as 1/γ0,p. ticipating supporting photons can compensate for a low energy contribution eV of the bias voltage by `s~ωs, thus shifting the onset of the exponential increase of F to a Figure3(b) shows results of reflection measurements at lower bias voltage. However, the corresponding relaxation different probe frequencies, bias voltages, and supporting- |`s| rates are suppressed by a typically small scaling factor ρs mode drive strengths. For an increasing bias voltage 2 (s) |` | or drive power, we observe broadening of the resonance arising from the transition matrix elements M ∝ ρ s kl s dip in frequency, indicating the activation of the QCR. (see AppendixD). On the other hand, the matrix elements Here, the broadening of the full width of the dip at half 2 (s) |`s| maximum corresponds to stronger total coupling strength depend on the occupation as M ∝ k for k 1/ρs, kl γ = γ +γ +γ of the system to its environments. which allows to sequentially activate the multiphoton tot,p tr,p T,p 0,p Remarkably, the resonance frequency of the primary mode, processes by increasing the occupation of the supporting located near the minimum reflection amplitude, exhibits mode until the rates saturate at ρ k ≈ 1, beyond which s clearly non-monotonic behavior which is repeated in the the matrix elements diminish as k−1/2. Thus, the dissipa- vicinity of each bias voltage corresponding to an onset of tion of the primary mode can be controlled by the drive a different multiphoton tunnelling process. Interestingly, strength of the supporting mode. this shift seems to be increased at high rf powers where it may persists even in the absence of voltage bias, but C. Reflection measurement Fig.3(b) does not provide clear enough evidence for these conclusions. To experimentally demonstrate the above-discussed phenomena, we probe the fundamental mode of the res- Thus for a detailed analysis, we extract ω , γ , γ , onator in a standard microwave reflection measurement p T,p tr,p and γ0,p by fitting the reflection coefficient to data cor- as illustrated in Fig.1(b). Ideally, the voltage reflection responding to Fig.3(b) (see AppendixC and Fig.8 coefficient of a weak probe signal at the angular frequency therein). For the corresponding theoretical model, in- ω p,probe is given by (see AppendixC) cluding equation (1), we map the rf drive power to the γ − γ − γ + 2i(ω − ω ) average supporting-mode occupation n¯s as described in Γ = tr,p T,p 0,p p,probe p , (2) γ + γ + γ − 2i(ω − ω ) AppendixE and Fig.9 therein. In addition to the trans- tr,p T,p 0,p p,probe p mission line and excess losses, we take into account for where γtr,p is the coupling strength between the resonator both modes the loss of rf photons owing to the different and the transmission line, and γ0,p is the damping rate types of photon-assisted electron-tunnelling events. 5

D. On-demand dissipation lation for each multi-photon-assisted electron tunnelling process, changing between −10 MHz to 22 MHz with Figure4(a) shows the measured environmental coupling the bias voltage. Importantly, under pure rf control the − . strength γT,p as a function of the bias voltage at differ- observed effective Lamb shift lies between 0 6 MHz to ent drive powers for the supporting mode. As discussed 0 MHz. Thus in the case of a purely rf-controlled QCR, above, with the addition of the rf excitation, the photon- fluctuations in the control parameters of the QCR may assisted electron-tunnelling events can absorb photons lead to less spurious dephasing of the coupled quantum from the supporting mode in addition to the primary- system than for a dc-controlled QCR. Furthermore, the mode photons. As a result of these additional energy rf-controlled effective Lamb shift appears as a non-linear quanta, the coupling strength γT,p exhibits in Fig.4(a) interaction between the two resonator modes which is a an exponential rise at a reduced bias voltage. We observe key constituent for using otherwise linear resonators as that with a sufficiently strong rf drive, the bias voltage qubits. is not even necessary for turning on the QCR-induced Next, we validate our claim that the observed frequency dissipation, i.e., the device is purely rf-controllable. This shift is in fact the effective Lamb shift by comparing our is beneficial in practical applications since undesired low- experimental observations with a first-principles theoreti- frequency noise can be significantly reduced by omission cal model which we derive in AppendixF. For the dynamic of dc connections to the sample. Since the rf control has a effective Lamb shift, which is the full environment-induced similar energy-providing effect to that of the bias voltage, frequency shift with the static, zero-frequency component it does not essentially change the maximum achievable subtracted, we obtain coupling strength. Consequently, the coupling strength ∞ Z γT,p is tunable by three orders of magnitude, roughly dω γT,p(ω) γT,p(ω) γT,p(ω) ωL = −PV 0 + 0 − 2 , (3) from 2π × 10 kHz to 2π × 10 MHz. 2π ω − ωp ω + ωp ω 0 Although for γT,p γtr,p, γ0,p the measured reflection coefficient is only weakly dependent on the quantity of where PV is the Cauchy principal value integration and interest, the results are in substantial agreement with the coupling strength γT,p(ω) is given by equation (1) such our theoretical model of the rf QCR as demonstrated in that V and n¯s are considered independent of ω = ωp. In Fig.4(a). However, note that the supporting-mode atten- addition, the fundamental frequency of a harmonic oscil- uation has been used as a fitting parameter as well as a 2 lator undergoes a classical damping shift of ∼ γT,p/(8ωp), single electron temperature TN that is independent of the although in our system this is only of the order of 10 kHz. bias voltage or drive power (see AppendicesB andC). We Furthermore, we observe a static shift of −µγT,p/π with expect a more comprehensive study of the electron tem- a proportionality constant µ = 0.64 at the two lowest perature to explain in the future most of the differences supporting-mode drive strengths. We attribute this shift between the model and the measurements, such as the to an effective elongation of the resonator mode [22], aris- interesting behavior observed at high powers where the ing from the exponentially increased tunnelling rate which oscillations of the experimentally observed damping are leads to a decreased effective junction impedance, and are much more pronounced than those of the model. This hence increased current flow through the junction. We difference cannot be explained by typical experimental do not claim the existence of a static shift at the higher noise which tends to smoothen such oscillations. drive strengths owing to the experimental uncertainties. The model describes well the experimental observations, as shown in Fig.4(b), except for the large measured ef- E. Photon-number-dependent effective Lamb shift fective Lamb shift at the highest rf drive strengths and bias voltages. This discrepancy is consistent with the Figure4(b) shows the observed shift in the primary- differences between the measured and the modeled cou- 0 mode frequency ωL = ωp − ωp as a function of the bias pling strengths visible in Fig.4(a). These differences are 0 voltage and of the rf drive strength, where ωp is the fre- partly explained by multiphoton events involving only quency for the QCR turned off. We observe that the the supporting mode, a phenomenon which is possible to shift strongly depends on the rf drive strength, and thus include in our model, however not considered here. on the supporting-mode photon number. Since the lin- In the reminder of this section, let us discuss possible ear resonator mode exhibits no ac Stark shift by the alternative explanations for the observed frequency shift. environment, we conclude that the observed shift is an Fortunately, we find that such alternatives are not in line effective Lamb shift that depends on the photon number with the experimental data, thus supporting the validity in the engineered hybrid environment through the photon of the above-given interpretation. number dependence of the coupling strength of Eq. (1). The frequency shift arising from the transformation of The fermionic part of the hybrid environment, i.e., elec- linearly coupled harmonic oscillators into the correspond- tron tunnelling, is required to mediate the interaction ing normal modes is equal for the quantum and classical between the bosonic system and the bosonic part of the models. Occasionally, this type of a contribution, referred environment. to as normal-mode splitting, is not considered to be a The effective Lamb shift in Fig.4(b) exhibits an oscil- part of the Lamb shift [30]. In our discussion above, we 6

FIG. 4. Coupling strength and the effective Lamb shift. Measured (dots) and modelled (solid lines) coupling strength γT,p (a) and effective Lamb shift ωL (b) as functions of the bias voltage at the indicated supporting-mode drive powers Ps in dBm. The estimated supporting-mode steady-state average photon number n¯s(V = 0) is also shown. Top axis shows the energy needed from the supporting-mode photons to activate the QCR for the primary mode. In (a), the external coupling strength γtr,p is indicated with the dashed line and the excess coupling strength γ0,p with the dash-dotted line. The shading denotes the 1σ confidence intervals of the extracted parameters as discussed in AppendixC. did not explicitly differentiate between these. However, since we measure the effective Lamb shift with respect to the zero-bias zero-drive-power case, our result is free from any normal-mode splitting that may occur owing to the presence of the modes of the resonator. Furthermore, the measured effective Lamb shift is mediated by electron tunneling with no direct coupling of the primary mode to additional physical bosonic modes. Thus, we do not consider the classical normal-mode splitting to significantly contribute to our observations. FIG. 5. Self-Kerr. (a) Experimentally measured reflection magnitude (not normalized) as a function of the probe To study possible contributions of the typical ac Stark frequency and the probe power. The data are taken at shift on our results, we have made an attempt to mea- Ps = −92.4 dBm and V = 0.095 mV. At the resonance, sure in a standard microwave reflection experiment the Pp = −116 dBm corresponds to an average primary-mode pho- strength of the self-Kerr effect in the primary mode as a ton occupation n¯p of the order of 1. (b) Resonance frequency function of probe frequency and power as shown in Fig.5. extracted from the reflection coefficient (dots) as a function of We find no self-Kerr effect within the experimental uncer- the occupation n¯p. The dashed line indicates the theoretically tainty of well below 0.1 MHz for hundred photons in the predicted Lamb-shifted resonance frequency. The slope of primary mode. In Fig.4(b) in contrast, hundred photons the linear fit (red) with an intercept ωint/2π = 8.82230 GHz in the supporting mode lead to a megahertz-level shift at (1σ uncertainty 0.025 MHz) yields estimated self-Kerr coef- the considered bias point. Since both modes are coupled ficient −0.1 kHz (1σ uncertainty 0.25 kHz). The intercept in a similar fashion to the NIS junction, the strengths of of the linear fit bears an additional systematic uncertainty of approximately 0.2 MHz which however does not affect the any self-Kerr and cross-Kerr effects promoted by the junc- slope. tion are expected to be roughly equal. Consequently, our experimental data does not support the cross-Kerr effect as an origin of the observed frequency shifts. In addition, any ac Stark shift arising from the cross-Kerr terms is is not measurably dependent on the effective population monotonic as a function of photon number, ∝ n¯s, whereas of the QCR environment either, and hence we may also the observed shift in Fig.4(b) between 0 .05–0.1 mV is rule out the possibility that the observed effective Lamb not. Thus, we conclude that the ac Stark shift of the shift could be modelled as an ac Stark shift arising from primary mode, arising from the photon occupation of the the effective electromagnetic environment induced by the secondary mode, has likely a negligible contribution to QCR. the measured frequency shifts. Furthermore, in this bias Furthermore, since the resistance of the NIS junction range the effective temperature of the QCR environment is non-linear, it may give rise to a classical-like frequency is the largest, see Fig.6, whereas the measured frequency shift which changes as a function of the bias voltage and shifts are on the smallest end. Thus, the observed shift drive power of the supporting mode. This consideration 7

ACKNOWLEDGMENTS

This research was financially supported by the Euro- pean Research Council under grant no. 681311 (QUESS) and Marie Sk lodowska-Curie grant no. 795159; by the Academy of Finland under its Centres of Excellence Pro- gram grant nos. 312300, 336810 and 312059 and grant nos. 265675, 305237, 305306, 308161, 314302, 316551, 316619, and 318937; and by the Alfred Kordelin Founda- tion, the Emil Aaltonen Foundation, the Vilho, Yrjöand Kalle VäisäläFoundation, the Jane and Aatos Erkko Foun- dation, and the Technology Industries of Finland Centen- FIG. 6. Characteristics of the effective electromagnetic nial Foundation. We thank the provision of facilities and environment formed by the QCR. Temperature (blue line, left technical support by Aalto University at OtaNano – Mi- axis) and the corresponding thermal occupation (orange line, cronova Nanofabrication Centre. right axis) of the QCR environment as functions of the bias voltage at the supporting-mode drive power Ps = −118.4 dBm. Top axis shows the energy needed from the supporting-mode Appendix A: Sample fabrication photons to activate the QCR for the primary mode.

We fabricate the sample, see Fig.7, on a high-purity includes the case of frequency mixing between the two 500-µm-thick silicon wafer passivated with a thermally- modes, the frequencies of which differ roughly by a factor grown silicon oxide layer of 300-nm thickness. The copla- of two. However, since the differential conductance of an nar waveguide resonator is patterned with photolithogra- NIS junction has less than three minima and maxima for phy and reactive ion etching in a 200-nm-thick sputtered positive bias voltages and we observe three local maxima niobium layer. A 50-nm-thick dielectric layer of Al2O3 is and minima of the frequency shift in Fig.4(b) such an atomic-layer deposited at 200 °C on the entire wafer to op- explanation for the phenomenon responsible of our results erate as the insulator in the parallel plate capacitors. The is ruled out. NIS junctions are defined with electron-beam lithography Interestingly, our findings are qualitatively supported and two-angle evaporation as follows: First, a 20-nm- by a rather general theoretical approach to the Lamb shift thick superconducting Al layer is evaporated, followed by from hybrid environments with initial results derived in in-situ oxidation to form the tunnel barriers. Second, a 20- AppendixG. We aim to develop this theory further in nm-thick normal-metal Cu layer is evaporated. Further the future. fabrication details are discussed in ref. [33].

III. DISCUSSION Appendix B: Measurements

In conclusion, we demonstrated and quantitatively mod- We cool the sample down to a base temperature of elled dc and rf control of the coupling strength between 10 mK in a commercial dilution refrigerator. The sample a resonator mode and its electromagnetic environment is attached to a gold-plated copper sample holder with formed by photon-assisted electron tunnelling in NIS junc- a printed circuit board, to which the sample is bonded tions. We found the dissipation to be tunable by three with aluminum wires. orders of magnitude using only rf driving. Peculiarly, We measure the reflection coefficient of the sample with we observed a photon-number-dependent effective Lamb a vector network analyzer (VNA) at frequencies close to shift of the linear resonator mode which oscillates as a that of the primary mode of the CPW resonator. We function of the bias voltage owing to the staggered onset drive the supporting mode with a microwave signal gen- of the multi-photon absorption processes. Each process erator through the same port as we use for the VNA, as strengthens the coupling of the bosonic mode to the en- illustrated in Fig.1b. The VNA tone is attenuated to the vironment, resulting in the oscillation in the coupling single-photon regime before it is introduced to the sample. strength and effective Lamb shift. Importantly, using The rf drive power for the supporting mode is controlled tunable rf power at vanishing bias voltage the dissipation by three orders of magnitude. The rf tone attenuation of strength can be controlled with almost vanishing effective the primary mode, −103 dB, is obtained by separately Lamb shift. Our device concept provides opportunities for measuring the room temperature components and using rf-controlled rapid on-demand initialization of quantum the manufacturer specifications for the cryogenic com- circuits, such as qubits [31, 32]. We envision the rf QCR ponents. For the supporting mode, the frequency range to become an important component in the emerging field is outside that specified for some cryogenic components, of quantum computing and technology. and consequently we use −105 dB obtained as a fitting 8

FIG. 7. Sample images. (a) Scanning electron micrograph of the superconductor–insulator–normal-metal tunnel junctions of a quantum-circuit refrigerator. The scale bar denotes 5 µm. (b) Optical image of a device similar to the the one used in this study. The tunnel junctions (red box) are connected to the coplanar-waveguide resonator (winding line), which is excited through the transmission line (broadening line). The dc junction bias is applied through RC ﬁlters between the bonding pads and the junctions. The test structures shown at the top middle of the image are not relevant for the operation of the device. The width of the ﬁeld of view is 7.5 mm.

a double Lorentzian of the form r = Γ(V,Ps) , where each Γ(0,P0) reﬂection coeﬃcient Γ is given by

2γ − r [γ + γ + γ − 2i(ω − ω )] 100 Γ = tr,p 0 tr,p T,p 0,p p,probe p , γtr,p + γT,p + γ0,p − 2i(ωp,probe − ωp) (C1) where r0 is a complex-valued Fano factor [35] such that |r0| = 1. This procedure is carried out for each voltage V and rf drive strength Ps, yielding a background-subtracted -10 -5 0 5 10 15 0 off-state reflection trace Γ (0,P0) which is determined by equation (C1) with the off-state fit parameters averaged over those obtained from the fits for different V FIG. 8. Reflection data analysis. Reflection magnitude |Γ| as and Ps. We retrieve the background-corrected traces as a function of the probe frequency ωp,probe at two different bias Γ0(V,P ) = rΓ0(0,P ) and fit equation (C1) to the ob- voltages and supporting mode powers: V = 0.08 mV,Ps = s 0 −90.8 dBm (red) and V0 = 0 mV,P0 = −118.4 dBm (yellow). tain the results for the primary-mode frequency ωp, the Fits to several fractions of two reflection magnitudes such as coupling strength γT,p, the external coupling strength r = |Γ(V,Ps)/Γ(V0,P0)| (blue) are used to extract an averaged γtr,p, and the excess coupling strength γ0,p used in this 0 off-state reflection magnitude Γ (V0,P0). Example of back- manuscript. ground subtracted reflection magnitude (green) is obtained as Ideally, in the absence of the Fano correction (r0 = 1), 0 0 |Γ (V,Ps)| = |rΓ (V0,P0)|. the reflection coefficient vanishes at the critical points where the impedance of the transmission line and of the resonator are matched at ωp,probe = ωp and γtr,p = parameter. The superconductor–insulator–normal-metal– γT,p + γ0,p. For these points, the full width of the dip, insulator–superconductor (SINIS) junction is biased with 2γtr,p, yields accurately the coupling strength to the trans- a ground-isolated room temperature voltage source. mission line. However, a non-vanishing Fano correction may shift the dip and make it shallower, although in our measurements the correction is small. Note that the Appendix C: Data analysis critical points can also be identified from the phase of reflection coefficient as it exhibits a full 2π phase winding We subtract the background in the measured VNA fre- about the critical points. The damping rate owing to ex- quency traces as described in ref. [34] and illustrated in cess losses γ0,p is extracted at zero voltage and minimum Fig.8. We discuss the method here because of the addi- supporting-mode power where the QCR is effectively de- tional tuning parameter, the supporting mode excitation. coupled from the rest of the circuit, i.e., γT,p is negligible. Namely, we divide the raw measured reflection coefficient Note that the measurements for γT,p γtr,p, γ0,p are Γ(V,Ps) by the off-state reflection trace Γ(0,P0) and fit challenging since the measured quantity is essentially a re- 9 sult of a subtraction of several other large quantities, thus ture individually for each measurement point would result amplifying the relative uncertainty of the result. Conse- in an artificially good match between the theory and data. quently, our results may be unreliable at such a regime, Expanding on this procedure, the device may be used as which is also reflected as relatively large experimental a thermometer. In this case, the device parameters may uncertainty bounds obtained as described below. be fitted at the slope of γT,p as a function of voltage at We obtain the 1σ confidence intervals of each ex- the lowest supporting-mode drive strength. Then, the tracted parameter of equation (C1) individually. In this rest of the data may be fitted with TN as the sole free method [22], we vary the parameter while the other pa- parameter. Further details of parameter extraction can rameters correspond to the optimal least-squares fit and be found in ref. [22]. calculate the resulting resonance point given by equation (C1). The 1σ confidence interval of the parameter is determined by the condition that the distance of the Appendix D: Photon-assisted electron tunnelling calculated resonance point from that of the least-squares fit in the complex plane is smaller or equal to the root- Following the approach of ref. [29] for a single mode, mean-square fit error of the reflection coefficient. we derive the coupling strength between the environment The quasiparticle temperature of the superconducting and the resonator based on the lumped-element circuit leads and the electron temperature of the normal-metal model given in Fig.2 for the two modes. island are higher than the base temperature of the re- We begin by assuming that the two modes are far off frigerator, for example, owing to heat leakage through resonance with each other, and thus the effect of their the radiation shields and to heating from the input sig- capacitive interaction is negligible to the lowest order. nals. Note that such low heating power does not affect Our quantum-mechanical circuit is thus described by the the temperature of the macroscopic dilution refrigerator Hamiltonian [36, 37] as much as the temperature of the small normal-metal island which is only weakly thermally coupled to the di- 2 2 (Qˆp + αpQˆN) (Qˆs + αsQˆN) lution refrigerator. The temperature affects the Fermi Hˆ = + C0 C0 distributions and determines the slope of the exponential 2 p 2 s growth in the coupling strength as a function of the bias ˆ2 ˆ 2 ˆ 2 QN Φp Φs voltage for the onset of each multiphoton process, but + 0 + + , (D1) 2C 2Lp 2Ls not the coupling strength, to which each photon-assisted N electron tunnelling process saturates to. Although in gen- where Qˆ , Qˆ , and Qˆ are the conjugate charge opera- eral the temperature is also affected by the amount and p s N tors of the primary and secondary resonator modes and type of quasiparticle tunneling processes, we observed the normal-metal island, respectively, Φˆ and Φˆ are the that our data is fairly well described by a single bias and p s respective node flux operators, and the capacitance ratios power-independent quasiparticle temperature. This is and the renormalized capacitances are given by likely because the temperature has a noticeable broadening effect only on steep slopes, which occur only at low Cc,p(Cc,s + Cs) input powers and at short voltage ranges. Thus, although αp = , (D2) C (C + C ) + C (C + C + C ) the temperature is in principle input power and voltage c,s c,p Σ s c,p c,s Σ dependent in the measurement range [33], we use for simplicity a single value of the quasiparticle temperature TN 0 CsCc,s in Table 1 for all measurement points, obtained by the Cp = Cp + αp CΣ + , (D3) best fit of the theoretical model to the curve in Fig.4 with Cc,s + Cs the lowest supporting-mode power. Fitting the tempera-

0 Cc,s(Cp + Cc,p)Cs + (Cp + Cc,p)(Cs + Cc,s)CΣ + Cc,p(Cs + Cc,s)Cp CN = , (D4) −Cc,s(Cp + Cc,p)αs + (Cp + Cc,p)(Cs + Cc,s) − Cc,p(Cs + Cc,s)αp

where CΣ = Cm+Cj is the total capacitance of the normal- by the transition matrix elements 0 metal island and αs and Cs are obtained by swapping the 2 (p/s) subscripts s and p in the above equations. Mmm0 m0 sgn(`) 2 −ρp/s |`| ! |`| = e ρ L 0 (ρ ) , (D5) The tunnelling events are considered as weak perturba- p/s m! min(m,m ) p/s tions inducing transitions between the eigenstates |mp/si πα2 0 p/s 0 h |m i where ρp/s = , ` = m − m , RK = 2 is the von and p/s of each resonator mode and are characterized ωp/sCp/sRK e 10

|`| Klitzing constant, and Lmin(m,m0)(ρp/s) is the generalized perconductor and of the normal metal, respectively. The Laguerre polynomial [38]. function nS is the normalized quasiparticle density of We assume the two NIS junctions to be identical, the states in the superconductor [39] electrodes to be at equal temperature TN, and the charging 2 0 ( ) energy EN = e /(2CN) of the normal-metal island to ε + iγD∆ nS(ε) = Re , (D8) be much smaller than other relevant energy scales, i.e., p 2 2 (ε + iγD∆) − ∆ EN ~ωp/s, ∆, kBTN, where ∆ is the Bardeen–Cooper– Schrieffer gap parameter. Fermi’s golden rule describes where γ is the Dynes parameter. the transition rates between the resonator number states D The interaction parameter of the primary resonator ρ as p is typically small, and thus at low powers, transitions between adjacent primary-mode states Γ(m ,m ),(m ±1,m0 ) Γ(m ,m ),(m0 ,m0 )(V ) p s p s p s p s are dominant, see equation (D5). We obtain the cou- 2 2 (p) (s) 2RK X pling strengths of the individual resonators by tracing = M 0 M 0 mpmp msms RT out the other resonator as, for example, for the primary τ=±1 k P −n¯s n¯s mode Γ 0 = P Γ 0 , where P = e F (τeV + `p~ωp + `s~ωs − EN), (D6) emm k,l k (m,k),(m ,l) k k! is the likelihood that the supporting mode Fock state where RT is the tunnelling resistance. Here, we denote |ki is occupied due to a classical drive and n¯s is the by F (E) the normalized forward tunnelling rate at the mean supporting-mode photon number, and by noting bias energy E the relations between the transition rates and the coupling strength, γT,p = Γe10 − Γe01, and the effective mode 1 Z h i−1 ~ωp Γe10 F (E) = dε nS (ε)[1 − fS(ε)]fN(ε − E), (D7) temperature, TT,p = k ln . Consequently, h B Γe01 the effective electromagnetic environment of the primary where fS and fN are the Fermi distributions of the su- mode is characterized by its coupling strength

Z 2 2 p X (s) X γT,p(V, n¯s) = 2παp Pk(¯ns) Mkl `pF (τeV + `p~ωp + `s~ωs − EN), (D9) RT k,l `p,τ=±1 where `s = k − l and the eﬀective mode temperature

  2 −1 P (s) P Pk M F (τeV + ωp + `s ωs − EN) ωp k,l kl τ=±1 ~ ~ T V, n ~    . T,p( ¯s) = ln  2  (D10) kB P (s) P k,l Pk Mkl τ=±1 F (τeV − ~ωp + `s~ωs − EN

The average thermal occupation of the environ- the annihilation and creation operators of the supporting ment at the temperature TT,p is given by NT,p = mode. The combined eﬀect of the drive and dissipation 1/{exp [~ωp/(kBTT,p)] − 1}. results in a steady-state occupation of In addition to the coupling strength, we have derived equation (3) for the eﬀective Lamb shift using second- 2 2 order perturbation theory for the eigenfrequencies of |Ωs| 2ωs 2 n¯s = 2 2 = 2 2 PsCg,sZtr,sZs, (E2) the primary mode. This derivation is presented in Ap- γtot,s + ∆s π~(γtot,s + ∆s ) pendixF.

where ∆s is the detuning of the rf drive, Zs is the charac- Appendix E: Mapping of input power to occupation teristic impedance of the supporting mode and Ps is the rf drive strength at the input capacitor of size Cg,s. Here, The driven supporting mode is modelled by the rf drive γtot,s = γT,s + γ0,s + γtr,s is the total coupling strength of Hamiltonian in the rotating frame the supporting mode and consists of the coupling of the mode to the QCR γT,s, to external sources γtr,s, and to ˆ ∗ † Hdrive = ~(Ωsaˆ + Ωs aˆ ), (E1) excess sources γ0,s. We assume the damping rate due to excess sources γ0,s to have a minor eﬀect here since the where Ωs is the complex-valued rf drive amplitude as external coupling is much higher than it. The coupling the resulting Rabi angular frequency and aˆ and aˆ† are strength of the supporting mode to the transmission line 11

Appendix F: Derivation of the eﬀective Lamb shift arising from multiphoton-assisted electron tunnelling events

Here, we derive the dynamic effective Lamb shift of the primary mode, Eq. (3), which originates from the photon- assisted electron tunnelling at the two capacitively coupled FIG. 9. Extraction of the supporting-mode occupation. The superconductor–insulator–normal-metal tunnel junctions. primary mode is driven at strength P , which due to the total p We take into account also the tunnelling events induced coupling strength γtot,p results in a steady-state occupation n¯p. This occupation affects the coupling strength of the supporting by an rf drive on the supporting mode. See the main text mode to the QCR, γT,s, which together with the external γtr,s for the definitions of the special terms. The derivation and excess coupling γ0,s of the mode is the total coupling begins by following the approach of refs. [22, 29, 40] but to strength γtot,s. The steady-state occupation n¯s is obtained the best of our knowledge has not been presented before from γtot,s and the supporting mode drive strength Ps using in the literature in this case of two bosonic modes coupled Eq. (E2). to the QCR such that one of the modes is traced out.

Let us denote by |ηi = |QN, mp, ms, `, ki = |QN, mp, msi⊗|`, ki the eigenstate of the combined unper- is calculated as [29] turbed system composed of the electrical circuit degrees of freedom |Q , m , m i and those of the quasiparticles in Z ω3 N p s s s the normal metal and superconductor of the QCR |`, ki. γtr,s = 2 2 , (E3) Ztr,s ωs + ωRC,s Owing to the pertubation intorduced by the tunnelling ˆ Hamiltonian HT, the energy-level shift ~δη of |ηi, where ω / Z C where RC,s = 1 ( tr,s g,s) is the corresponding angular ~ is the reduced Planck constant, is given by the second- frequency. The resonator frequency ωs/(2π) originates order time-independent perturbation theory as [22] from 2 D E 2 0 Zs ωs ωRC,s 0 ˆ ω ≈ ω − , (E4) η HT η s s Z ω2 ω2 0 X 2 tr,s s + RC,s ~δη = Eη − Eη = − , (F1) Eη0 − Eη η06=η 0 where ωs/(2π) is the renormalized resonator frequency which is experimentally measured and employed as the where the total energy is E = E Q2 + ω0m + ω m + rf drive frequency. The coupling strength between the η N N ~ p p ~ s s ε + . Here, the integer Q is the eigenvalue of charge supporting mode and the QCR, γ , is obtained from ` k N T,s for the normal-metal QCR island, ω0 is the bare angular equation (D9) by exchanging the roles of the primary p frequency of the primary mode, ω is the angular fre- and supporting modes. Furthermore, the steady-state s quency of the supporting mode, and m denotes the occupation of the primary mode n¯ is calculated with p/s p number of photons in the primary or the supporting equation (E2) by substituting the measured values of mode. Furthermore, the charging energy of the QCR γ , the primary-mode drive strength P , and the de- tot,p p island is E = e2/(2C0 ), where C0 is the renormalized tuning of the primary-mode drive ∆ . The procedure is N N N p capacitance of the island provided by Eq. (D4). The illustrated in Fig.9. quasiparticles of the normal-metal and superconducting In the future, we intend to improve the model by itera- electrodes have energies ε and , respectively, where tion of the above procedure and alternatively, by replacing ` k ` and k index these and the corresponding eigenstates it with solving throughout this document implicitly including the spin ( degree of freedom. The quasiparticle tunnelling events 0 = −i∆ (¯n )¯α − γtot,s(¯np) α¯ + iΩ s p s 2 s s between the normal-metal and superconducting leads are γtot,p(¯ns) (E5) 0 = −i∆p(¯ns)¯αp − 2 α¯p + iΩp, described by the tunnelling Hamiltonian [40] where α¯s/p is the steady-state amplitude of the coherent e e X † −i φˆN ∗ † i φˆN 2 Hˆ = T dˆ cˆ e ~ + T dˆ cˆ e ~ , (F2) state in the supporting/primary mode obeying |α¯s/p| = T `k ` k `k ` k `k n¯s/p. In addition to the driven occupation given by Eq. (E2), the resonator mode has a thermal population where T`k is a tunnelling matrix element, the annihilation 1/ {exp[~ωs/(kBT )] − 1}. We neglect this as it is orders operators of the normal-metal and of the superconducting of magnitude smaller than the driven occupation in the leads are denoted by dˆ` and cˆk, respectively, φˆN is the temperature scales TN,TT,s of the experiment [29, 33], node flux operator of the normal-metal island, and e is although a more complex sample capable of a precise tem- the elementary charge. These events cause weak pertur- perature measurement is needed to fully rule this effect bations to the coupled electric circuit shown in Fig.2, out. described by the core Hamiltonian which we represent 12 using its eigenstate decomposition as Here, we denote the first term owing to elastic transi- el tions as ω 0 and the second term owing to the ∞ ∞ L,p,ms,ms X X 2 0 ph Hˆ E Q ω m ω m primary-mode-assisted transitions as ω 0 . Note = N N + ~ p p + ~ s s L,p,ms,ms el QN=−∞ mp,ms=0 that the elastic transition term is simply ω 0 = L,p,ms,ms ph 0 × |QN, mp, msi hQN, mp, ms| , (F3) − limω0→0 ω 0 ω and it thus suffices to simplify p L,p,ms,ms p the primary-mode-assisted part. where the eigenstates of the core Hamiltonian are We express the matrix elements of the quasiparticle obtained from the eigenstates of the charge |QNi D E transitions `0k0|Θˆ |`k with the normalized forward and from the Fock states |mpi and |msi of the primary and supporting modes as |QN, mp, msi = quasiparticle tunnelling rate F given in Eq. (D7). Here, e e we assume that the tunnelling matrix elements are ap- exp −iαpQN Φˆ p − iαsQN Φˆ s |QNi ⊗ |mpi ⊗ |msi, h h proximately constant in the vicinity of the Fermi energies, ˆ where Φp/s are the flux operators and αp/s the capacitance the two SIN junctions are identical, and the electrodes ratios of the resonator modes as described by Eq. (D2). are at equal temperature. Consequently, we obtain (p/s) We use the transition matrix elements M 0 of the mm ph ω 0 electric circuit given in Eq. (D5) and expand equation (F1) L,p,ms,ms assuming that EN is the smallest relevant energy scale of ρ R 2 Z ∞ 0 p K (s) X E , ω , ω k T = − M 0 PV d the system N ∆ ~ p ~ s and B N . After tracing msms π RT −∞ out the charge degree of freedom of the normal-metal and τ,`p=±1 of the superconducting leads, we obtain an energy shift 0 F + τeV + `p~ωp + `s~ωs − EN for the Fock state |mpi corresponding to the transition 0 |msi → |msi in the supporting mode as 2 Z ∞ 2 Zp (s) X dω = − 2πα M 0 PV p msms δ˜ 0 RT 0 2π ~ mp,ms,ms τ,`p=±1 2 2 X X X (p) (s) `pF (τeV + `p ω + `s ωs − EN ) = − Mm m0 Mm m0 ~ ~ p p s s 0 + PV 0 0 0 ω − ω mp `,` k,k p Z ∞  D E 2 dω `pF (τeV + `p~ω + `s~ωs − EN) `0k0 ˆ † `k , (F6) Θ 0  0 2π ω + ωp  0 EN + `p~ωp + `s~ωs + E`0k0 − E`k where PV denotes the Cauchy principal value integration, D E 2  RK is the von Klitzing constant, RT is the tunnelling `0k0| ˆ |`k Θ resistance, and Zp is the characteristic impedance of the +  , (F4) 0  primary mode in its lumped-element description. Fur- EN + `p~ωp + `s~ωs + E`0k0 − E`k thermore, by tracing out the supporting resonator mode

with occupation probabilities Pms and expressing equa- ˆ P ˆ† where Θ = `k T`kd`cˆk and its Hermitian conjugate is tion (F6) in terms of the coupling strength γT,p of the the quasiparticle part of the tunnelling Hamiltonian and effective electromagnetic environment ` = m0 −m is the change in the mode occupations. p/s p/s p/s Z 2 2 p X (s) X The interaction parameter of the primary mode is small γT,p(V, ωp) =2παp Pms Mm m0 2 R s s (p) T 0 ms,ms τ,`p=±1 ρp ≈ 0.003, which allows us to expand M 0 up mpmp ` F (τeV + ` ω + ` ω − E ) , (F7) to the first order in ρp. The corresponding dynamic p p~ p s~ s N effective Lamb shift of the primary mode ωL,p,m ,m0 = s s we obtain the dynamic effective Lamb shift of the primary δ˜ 0 − δ˜ 0 mp+1,ms,ms mp,ms,ms is thus given by mode Z ∞ ω 0 dω γT,p(V, ω) γT,p(V, ω) L,p,ms,ms 0 ωL,p V, ωp = −PV 0 + 0 ρ 2 0 2π ω − ωp ω + ωp p X X X (s) = Mm m0 s s γ V, ω ~ 0 0 T,p( ) `,` k,k `p=±1 −2 , (F8) ω  D E 2 D E 2 `0k0 Θˆ † `k + `0k0|Θˆ |`k  where the terms dependent on the primary mode fre-  0 EN + `s~ωs + E`0k0 − E`k quency ωp are contributions from the photon-assisted transitions and the last term arises from elastic tunnelling. D E 2 D E 2  0 0 ˆ † 0 0 ˆ Note that the effective Lamb shift of the primary mode ` k Θ `k + ` k |Θ|`k −  . (F5) depends on the quantum state of the supporting mode 0  EN + `p~ωp + `s~ωs + E`0k0 − E`k through the coupling strength γT,p and the occupation

probabilities Pms of the Fock states |msi. 13

b 0 −1 b = + (iωn − ωp) . Thus, if we consider iωn as a complex b variable, there is a pole at the oscillator frequency iωn = 0 FIG. 10. Diagrammatic form of the Dyson equation for the ωp. This property allows us to deﬁne the bath-shifted bosonic Green’s function. Here, the thin dashed lines corre- frequency of the mode as the position of the pole of the spond to the bare Green’s function of the bosonic mode, the Green’s function of the interacting system. The Green’s thin solid lines correspond to the bare Green’s functions of the function of the interacting system can be expanded as a fermionic bath, and the thick dashed lines correspond to the series of diagrams consisting of the Green’s functions of Green’s function of the bosonic mode with the interactions the non-interacting (bare) bosonic and fermionic modes. taken into account. The considered diagrams correspond to The main series of contributions which corresponds to the the Born approximation for the bosonic Green’s function. Born approximation for the Green’s function is shown in the Fig. 10. The Dyson equation corresponding to these diagrams reads as follows: Appendix G: Theory of the Lamb shift of a bosonic mode coupled to a fermionic bath −1 b−1 h (0) bi b D ωn = D ωn − Π(ωn), (G6)

We study a primary bosonic mode coupled to a bath +∞ 2 k T |γ 0 | of non-interacting fermions described by the following b B X X kk Π(ωn) = 3 f b f , (iω + iω − ξk)(iω − ξk0 ) Hamiltonian: ~ kk0 m=−∞ m n m (G7) ˆ 0ˆ†ˆ X † H =~ωpb b + ξkaˆkaˆk k f where ωm = π(2m + 1)kBT/~ are the fermionic Matsub- X ˆ † ∗ ˆ† † ara frequencies. Vanishing right side of the Dyson equa- + γkk0 baˆkaˆk0 + γkk0 b aˆk0 aˆk , (G1) kk0 tion (G6) correspond to the pole of the Green’s function, and, thus, its position yields the bath-shifted frequency. 0 where ωp is the bare frequency of the primary mode, ξk To proceed, let us introduce simplifying assumptions is the energy of k-th fermionic mode of the bath, and on the structure of the bath and the coupling parameters. the parameters γkk0 control coupling between the bosonic We assume that the energies of the fermionic modes of mode and the fermionic bath. We do not consider ex- the bath are uniformly distributed within the interval −µ citations of the bath arising from any coupling it to a to W − µ, where µ is the chemical potential and W is driven supporting mode. Instead we assume that the the width of the bath. Thus, we can replace summation P R W −µ whole system has some equilibrium but nonzero temper- over k with an integral k → νVbath −µ dξ, where ν ature T and study the temperature dependence of the is the density of states per the unit volume and Vbath is primary-mode frequency. This assumption allows us to the volume of the Fermi gas. In addition, we assume that use a finite-temperature diagrammatic technique for this absorption and emission of a boson couples all fermionic problem. 2 2 2 modes between each other and fix |γkk0 | = Γ /Vbath. We begin by defining an imaginary-time bosonic Consequently, we have a single parameter which controls 0 Green’s function D(τ, τ ) as the coupling strength. Under these assumptions the po- b n h io larization operator Π(ωn) can be evaluated as −βHˆ ˆ ˆ† 0 Tr e Tτ b(τ)b (τ ) D(τ, τ 0) = , (G2) W −µ −βHˆ 2 2 Z 2 b 2 Tr e b Γ ν (ε − W + µ) + (~ωn) Π(ωn) = ln 2 b 2 −1 2~ (ε + µ) + (~ωn) where β = (kBT ) , Tτ is the imaginary-time-ordering −µ operator and the operators ˆb(τ) and ˆb†(τ) are defined as ε tanh dε. (G8) 2kBT Hτˆ − Hτˆ ˆb(τ) = e ~ ˆbe ~ , (G3) Finally, the equation for the frequency shift can be ex- † Hτˆ † − Hτˆ ˆb (τ) = e ~ ˆb e ~ . (G4) pressed as

W −µ This Green’s function is defined in the interval 2 2 Z 0 Γ ν −~/(kBT ) < τ − τ < ~/(kBT ) and can be expanded ωL,p(T ) = 2~ into a Fourier series as −µ ∞ ( 2 0 2 ) 0 kBT X b −i(τ−τ 0)ωb (ε − W + µ) + [~(ωL,p + ωp) + i0] D τ, τ D ω e n , ( ) = ( n) (G5) ln 2 0 2 ~ n=−∞ (ε + µ) − [~(ωL,p + ωp) + i0] ε b tanh dε. (G9) where ωn = 2πnkBT/~ are bosonic Matsubara frequencies. 2kBT For the single bosonic mode uncoupled from the bath, the (0) b 0 Fourier expansion coefficients are equal to D (ωn) = In the limit Γν 1 and ~ωp, kBT µ, (W − µ) the 14 approximate solution of this equation can be found as The real part of ωL,p defines the bath-induced frequency shift of the bosonic mode whereas the imaginary part W −µ gives the dissipation rate. Γ2ν2 Z W − µ + ε ε ωL,p(T ) ≈ ln tanh dε ~ ε + µ kBT −µ 2 2 0 − iπΓ ν ωp 2 2 2 2 π Γ ν (kBT ) W ≈ωL,p(0) − 3~µ(W − µ) 2 2 0 − iπΓ ν ωp. (G10)

[1] R. R. Ernst, G. Bodenhausen, and A. Wokaun, Princi- dot, A. Maali, J. M. Raimond, and S. Haroche, From ples of Nuclear Magnetic Resonance in One and Two Lamb shift to light shifts: Vacuum and subphoton cavity Dimensions (Clarendon press, Oxford, 1987). fields measured by atomic phase sensitive detection, Phys. [2] M. Hämäläinen,R. Hari, R. J. Ilmoniemi, J. Knuutila, Rev. Lett. 72, 3339 (1994). and O. V. Lounasmaa, Magnetoencephalography—theory, [17] M. Marrocco, M. Weidinger, R. T. Sang, and H. Walther, instrumentation, and applications to noninvasive studies Quantum electrodynamic shifts of Rydberg energy levels of the working human brain, Rev. Mod. Phys. 65, 413 between parallel metal plates, Phys. Rev. Lett. 81, 5784 (1993). (1998). [3] T. D. Ladd, F. Jelezko, R. Laflamme, Y. Nakamura, [18] T. Rentrop, A. Trautmann, F. A. Olivares, F. Jendrze- C. Monroe, and J. L. O’Brien, Quantum computers, Na- jewski, A. Komnik, and M. K. Oberthaler, Observation of ture 464, 45 (2010). the phononic Lamb shift with a synthetic vacuum, Phys. [4] C. Neill, P. Roushan, K. Kechedzhi, S. Boixo, S. V. Isakov, Rev. X 6, 041041 (2016). V. Smelyanskiy, A. Megrant, B. Chiaro, A. Dunsworth, [19] A. Fragner, M. Göppl, J. Fink, M. Baur, R. Bianchetti, and K. Arya et al., A blueprint for demonstrating quan- P. Leek, A. Blais, and A. Wallraff, Resolving vacuum tum supremacy with superconducting qubits, Science 360, fluctuations in an electrical circuit by measuring the Lamb 195 (2018). shift, Science 322, 1357 (2008). [5] S. Rosenblum, P. Reinhold, M. Mirrahimi, L. Jiang, [20] F. Yoshihara, T. Fuse, Z. Ao, S. Ashhab, K. Kakuyanagi, L. Frunzio, and R. J. Schoelkopf, Fault-tolerant detection S. Saito, T. Aoki, K. Koshino, and K. Semba, Inversion of a quantum error, Science 361, 266 (2018). of qubit energy levels in qubit-oscillator circuits in the [6] F. Arute, K. Arya, R. Babbush, D. Bacon, J. Bardin, deep-strong-coupling regime, Phys. Rev. Lett. 120, 183601 R. Barends, R. Biswas, S. Boixo, F. Brandao, and D. Buell (2018). et al., Quantum supremacy using a programmable super- [21] M. Mirhosseini, E. Kim, V. S. Ferreira, M. Kalaee, conducting processor, Nature 574, 505 (2019). A. Sipahigil, A. J. Keller, and O. Painter, Superconducting [7] A. Blais, S. M. Girvin, and W. D. Oliver, Quantum in- metamaterials for waveguide quantum electrodynamics, formation processing and quantum optics with circuit Nat. Commun. 9, 3706 (2018). quantum electrodynamics, Nat. Phys. 16, 247 (2020). [22] M. Silveri, S. Masuda, V. Sevriuk, K. Y. Tan, M. Jenei, [8] C. H. Bennett and G. Brassard, Quantum cryptography: E. Hyyppä,F. Hassler, M. Partanen, J. Goetz, and R. E. Public key distribution and coin tossing, in Proc. IEEE Lake et al., Broadband Lamb shift in an engineered quan- Int. Conf. on Computers, Systems, and Signal Processing, tum system, Nat. Phys. 15, 533 (2019). Bangalore (1984) pp. 175–179. [23] K. Y. Tan, M. Partanen, R. E. Lake, J. Govenius, S. Ma- [9] A. K. Ekert, Quantum cryptography based on Bell’s the- suda, and M. Möttönen, Quantum-circuit refrigerator, orem, Phys. Rev. Lett. 67, 661 (1991). Nat. Commun. 8, 15189 (2017). [10] W. E. Lamb and R. C. Retherford, Fine structure of the [24] H. J. Carmichael, Statistical Methods in Quantum Op- hydrogen atom by a microwave method, Phys. Rev. 72, tics 1 (Springer, Berlin, 1999). 241 (1947). [25] P. Y. Wen, K.-T. Lin, A. F. Kockum, B. Suri, H. Ian, [11] H. A. Bethe, The electromagnetic shift of energy levels, J. C. Chen, S. Y. Mao, C. C. Chiu, P. Delsing, and Phys. Rev. 72, 339 (1947). F. Nori et al., Large collective Lamb shift of two distant [12] J. Bardeen, L. N. Cooper, and J. R. Schrieffer, Theory of superconducting artificial atoms, Phys. Rev. Lett. 123, superconductivity, Phys. Rev. 108, 1175 (1957). 233602 (2019). [13] H. B. G. Casimir and D. Polder, The influence of retar- [26] A. A. Clerk, K. W. Lehnert, P. Bertet, J. R. Petta, and dation on the London-van der Waals forces, Phys. Rev. Y. Nakamura, Hybrid quantum systems with circuit quan- 73, 360 (1948). tum electrodynamics, Nat. Phys. 16, 257 (2020). [14] I. Giaever, Energy gap in superconductors measured by [27] D. Lachance-Quirion, S. P. Wolski, Y. Tabuchi, S. Kono, electron tunneling, Phys. Rev. Lett. 5, 147 (1960). K. Usami, and Y. Nakamura, Entanglement-based single- [15] D. J. Heinzen and M. S. Feld, Vacuum radiative level shot detection of a single magnon with a superconducting shift and spontaneous-emission linewidth of an atom in qubit, Science 367, 425 (2020). an optical resonator, Phys. Rev. Lett. 59, 2623 (1987). [28] S. Barzanjeh, E. Redchenko, M. Peruzzo, M. Wulf, [16] M. Brune, P. Nussenzveig, F. Schmidt-Kaler, F. Bernar- D. Lewis, G. Arnold, and J. M. Fink, Stationary en- 15

tangled radiation from micromechanical motion, Nature Tan, M. Silveri, J. Goetz, M. Partanen, R. E. Lake, and 570, 480 (2019). L. Grönberg et al., Calibration of cryogenic amplifica- [29] M. Silveri, H. Grabert, S. Masuda, K. Y. Tan, and tion chains using normal-metal–insulator–superconductor M. Möttönen,Theory of quantum-circuit refrigeration junctions, Appl. Phys. Lett. 114, 192603 (2019). by photon-assisted electron tunneling, Phys. Rev. B 96, [35] U. Fano, Effects of configuration interaction on intensities 094524 (2017). and phase shifts, Phys. Rev. 124, 1866 (1961). [30] M. F. Gely, G. A. Steele, and D. Bothner, Nature of the [36] A. Wallraff, D. I. Schuster, A. Blais, L. Frunzio, R.-S. Lamb shift in weakly anharmonic atoms: From normal- Huang, J. Majer, S. Kumar, S. M. Girvin, and R. J. mode splitting to quantum fluctuations, Phys. Rev. A 98, Schoelkopf, Strong coupling of a single photon to a super- 053808 (2018). conducting qubit using circuit quantum electrodynamics, [31] V. Sevriuk, K. Y. Tan, E. Hyyppä,M. Silveri, M. Parta- Nature 431, 162 (2004). nen, M. Jenei, S. Masuda, J. Goetz, V. Vesterinen, and [37] A. Blais, R.-S. Huang, A. Wallraff, S. M. Girvin, and L. Grönberg et al., Fast control of dissipation in a su- R. J. Schoelkopf, Cavity quantum electrodynamics for perconducting resonator, Appl. Phys. Lett. 115, 082601 superconducting electrical circuits: An architecture for (2019). quantum computation, Phys. Rev. A 69, 062320 (2004). [32] H. Hsu, M. Silveri, A. Gunyhó,J. Goetz, G. Catelani, and [38] M. Abramowitz and I. A. Stegun, Handbook of Mathe- M. Möttönen,Tunable refrigerator for nonlinear quantum matical Functions (Dover, New York, 1972). electric circuits, Phys. Rev. B 101, 235422 (2020). [39] R. C. Dynes, V. Narayanamurti, and J. P. Garno, Direct [33] S. Masuda, K. Y. Tan, M. Partanen, R. E. Lake, J. Gove- measurement of quasiparticle-lifetime broadening in a nius, M. Silveri, H. Grabert, and M. Möttönen,Observa- strong-coupled superconductor, Phys. Rev. Lett. 41, 1509 tion of microwave absorption and emission from incoher- (1978). ent electron tunneling through a normal-metal-insulator- [40] G.-L. Ingold and Y. V. Nazarov, Charge tunneling rates superconductor junction, Sci. Rep. 8, 3966 (2018). in ultrasmall junctions, in Single Charge Tunneling: [34] E. Hyyppä,M. Jenei, S. Masuda, V. Sevriuk, K. Y. Coulomb Blockade Phenomena in Nanostructures (eds Grabert, H. & Devoret, M. H.) (Plenum, New York, 1992).