Quantum magnetic monopole condensate

M. C. Diamantini,1 C. A. Trugenberger,2 and V.M. Vinokur3 1NiPS Laboratory, INFN and Dipartimento di Fisica e Geologia, University of Perugia, via A. Pascoli, I-06100 Perugia, Italy 2SwissScientific Technologies SA, rue du Rhone 59, CH-1204 Geneva, Switzerland 3Terra Quantum AG, St. Gallerstrasse 16A, CH-9400 Rorschach, Switzerland Despite decades-long efforts, magnetic monopoles were never found as elementary particles. Monopoles and associated currents were directly measured in experiments and identified as topological quasiparticle excitations in emergent condensed matter systems. These monopoles and the related electric-magnetic symmetry were restricted to classical electrodynamics, with monopoles behaving as classical particles. Here we show that the electric-magnetic symmetry is most fundamental and extends to full quantum behavior. We demonstrate that at low temperatures magnetic monopoles can form a quantum Bose condensate dual to the charge Cooper pair condensate in superconductors. The monopole Bose condensate manifests as a superinsulating state with infinite resistance, dual to superconductivity. Monopole supercurrents result in the electric analog of the Meissner effect and lead to linear confinement of Cooper pairs by Polyakov electric strings in analogy to quarks in hadrons.

INTRODUCTION a b + – –

– + + – – + +

Maxwell’s equations in vacuum are symmetric under the dual- + – – +

– – + – – ity transformation E → B and B → −E (we use natural units +

c = 1, ~ = 1, ε = 0). Duality is preserved, provided that both + + 0 – + – – – + electric and magnetic sources (magnetic monopoles and mag- – – +

+ + – netic currents) are included [1]. Magnetic monopoles, while + – + +

– + elusive as elementary particles [2], exist in many materials in – + + the form of emergent quasiparticle excitations [3]. Magnetic – monopoles and associated currents were directly measured in experiments [4], confirming the predicted symmetry between electricity and magnetism. The existence of monopoles re- FIG. 1. Magnetic monopole states at low temperatures. a: quires that gauge fields are compact, implying, in turn, the Monopoles are confined into dipoles state by tension of vortices con- quantization of charge and Dirac strings [5] or a core with ad- necting them. b: As the vortex tension vanishes, the monopoles at ditional degrees of freedom to regularize the singularities of their endpoints become loose free particles that can condense. the vector potential [6,7]. The importance of electric-magnetic duality was first re- alized by Nambu [8], Mandelstam [9] and ’t Hooft [10], who analog of the Meissner effect and lead to linear confinement proposed that colour confinement in quantum chromodynam- of the Cooper pairs by Polyakov’s electric strings [14, 17, 18] ics (QCD) can be understood as a dual Meissner effect. In (dual to superconducting vortices) in analogy to quarks within the present context, electric-magnetic duality manifestly real- hadrons [19]. The monopole condensate realizes a 3D ver- izes the symmetry between Cooper pairs, which are Noether sion of superinsulators, that have been previously observed in charges, and magnetic monopoles, which are topological 2D superconducting films, and result from quantum tunneling solitons, and forms the foundation for the superconductor- events, or instantons [12–14, 17, 18]. arXiv:2007.02356v2 [hep-th] 17 Feb 2021 insulator transition [11] and the appearance of the superinsu- lating state [12–14]. Magnetic monopoles arise also as instan- tons in Josephson junction arrays (JJA) [15], which are easily RESULTS accessible experimental systems themselves and provide an exemplary model for superconducting films [16]. Magnetic monopoles in granular superconductors So far, monopoles in emergent condensed matter systems were treated as classical excitations. Here we show that the electric-magnetic symmetry is most fundamental and ex- To gain insight into the nature of a system that can harbor tends to the full quantum realm. We demonstrate here that Cooper pairs and magnetic monopoles simultaneously, we at low temperatures, magnetic monopoles form a quantum first reiterate that monopoles naturally emerge in 2D JJA [15] Bose condensate dual to the charge condensate in supercon- as instantons and provide the underlying mechanism of 2D su- ductors and generate a superinsulating state with infinite re- perinsulation as quantum tunneling events. To establish that sistance, dual to superconductivity [12, 13]. We show that JJA indeed offers a universal model describing superinsulation magnetic monopole supercurrents result in the direct electric in films, one recalls the early hypothesis that, in the vicin- 2 ity of the superconductor-insulator transition (SIT), the films Long-distance effective field theory acquire self-induced electronic granular texture and can be viewed as a set of superconducting granules immersed into an insulating matrix [20]. Employing the model of the 2D Having established that in the vicinity of the SIT an ensem- JJA [21] to treat the experimental data of [17, 22], perfectly ble of Cooper pairs acquires a granular structure, we con- confirmed this picture, and settled that granules have the typ- struct a Ginzburg-Landau-type effective field theory of such ical size of order of the superconducting coherence length, ξ, granular Cooper pair condensates. We focus on the London and are coupled by Josephson links. On the theory side, this limit, i.e. on long distances, much larger than the topological granular structure was derived in the framework of the gauge width scale so that vortices and monopoles appear as point- theory of the SIT in [23], conclusively establishing the JJA- like singularities. As a first step, following general princi- like texture in 2D systems experiencing the SIT. ples by Wilczek [25], we identify that the dominant interac- tions are the topological mutual statistics interaction setting As a next step, we generalize the reasoning of [23] onto 3D the Aharonov-Bohm phases between charges and vortices. A systems. This enables us to adopt the original gauge theory of standard description of the interaction part of the system’s JJA [12] for 3D superconductors and consider a general inho- Lagrangian is achieved by introducing two fictitious gauge mogeneous system of superconducting granules, i.e. bubbles fields, a vector field aµ and an antisymmetric pseudotensor of Cooper pair condensate, coupled by tunneling links. Then, gauge field bµν [26]: depending on the ratio of the strength of Josephson coupling 1 µναβ µ 1 µν and the Coulomb energy of the elemental excessive charge on LBI = bµν ∂αaβ + aµq + bµνm . (1) a single granule, the system can be either an insulator, super- 4π 2 conductor, or topological insulator [12, 23]. If the system is where qµ and mµν are the charge and vortex currents, respec- at the insulating side of the SIT, the global phase coherence tively. This so-called BF model [26] is topological, since it is absent and each condensate bubble is characterized by an is metric-independent. It is invariant under the usual gauge independent phase. The relevant degrees of freedom in such transformations aµ → aµ + ∂µξ and under the gauge trans- systems are single Cooper pairs that can tunnel from one is- formation of the second kind[26], under which the antisym- land to the next one, leaving behind a Cooper hole, and vor- metric tensor transforms as bµν → bµν + ∂µλν − ∂νλµ, a tices, resulting from non-trivial phase circulations over adja- vector field λµ becoming itself the gauge function. The BF cent granules. In a 2D system such a vortex would be a usual action for a model defined on a compact space endowed Josephson vortex, and in 3D such an elemental ’minimal’ vor- with the non-trivial topology, yields a ground state with the tex is similar to a pancake vortex in a layered cuprate [24]. In degeneracy reflecting, one-to-one, this topology and is re- the usual configuration, these pancake vortices aggregate on ferred to as topological order [27]. The coefficient 1/4π top of each other to make stacks, or chains. When the pan- of the first term in Eq. (1) ensures that the system does cake stack forms, the magnetic monopoles and antimonopoles not have such a topological order [28]. In turn, the topo- at the “bottom” and “top” of each of such a pancake “annihi- logical coupling between a vector and a pseudotensor in late” and one long vortex forms. Only the monopoles and anti- 3D, ensures the parity (ZP) and time-reversal (ZT) symme- monopoles at the very end of this configuration survive and, tries of the model. The field strength associated with aµ correspondingly, one has such monopoles and anti-monopoles is fµν=∂µaν − ∂νaµ, whereas bµν has the associated 3-tensor only at the surfaces of the sample [24]. When pancake vor- field strength hµνα=∂µbνα + ∂νbαµ + ∂αbµν. It can be eas- tices are ballistic and the layers are only weakly coupled, how- ily checked that this expression is invariant under the gauge ever, the vertical stacks can break in the middle, resulting in transformations of the second kind; it plays for bµν the same monopole and anti-monopole pairs joined by a shorter vor- role that the field strength fµν plays for the usual gauge field µ µ µναβ tex in the interior of the sample. As always, when there are aµ. The dual field strengths j =(1/2π)h =(1/4π) ∂νbαβ µν µν µναβ both dynamic charges and vortices in the spectrum, the latter and m = (1/2π) f˜ = (1/4π) fµν represent the con- acquires a topological gap [25]. The inverse of the gap sets served charge and vortex currents, respectively. For Cooper the spacial scale for the width of the vortices and radius of pairs, the charges are measured in integer units of the elemen- monopoles. tal charge of a Cooper pair, 2e. Accordingly, the magnetic charge of monopoles and vortices is measured in integer units On distances larger than the vortex width, long vortices of 2π/2e = π/e. If one defines a fundamental charge as a appear as quantized flux tube singularities that become charge of a single electron, e, then our magnetic monopoles unobservable because of the compactness of the U(1) gauge of strength π/e should be viewed, strictly speaking, as half- fields [1], and a neutral plasma of point magnetic monopoles monopoles. However, we always deal with the phases of mat- satisfying the Dirac quantization [1] condition forms. Re- ter where charge unit is 2e rather than e. Thus we shall call our markably, as we show below, the monopole plasma strongly objects the unit monopoles. This complies with condensed suppresses Cooper pair tunneling providing thus a mechanism matter notations, where a unit vortex carries magnetic flux that stabilizes the granular structure. quantum π/e, since it always appears in a Cooper pair con- densate, but differs from the field theory notation where π/e is 3 half a vortex. mining the phase structure. More specifically, when inverting The developed model describes what is known today as the Kalb-Ramond kernel, we will consider only the transverse a (simple, as opposed to strong) bosonic topological insula- components of the vortices and neglect their endpoints. tor [29] and is an intensely investigated state of matter. Open Rotating to Euclidean space-time we arrive at the action vortices emerging in this model carry magnetic monopole- 4 4 4 ν µν X ` ` ` antimonopole pairs with current m = ∂µm at their ends, S = f f − i a k b + h h f 2 µν µν 4π µ µαβ αβ g2 µνα µνα and the gauge symmetry of the second kind is broken. The x 4 12 state of the monopole ensemble is self-consistently harnessed 1 +i`a q + i`2 b m , (2) with vortex properties. Existence of the appreciable vortex µ µ 2 µν µν tension implies that vortices are short and taut and linearly where k is the lattice BF term [12], see Methods, f is a di- confine the monopoles into ‘small’ dipoles, as illustrated in µνα mensionless coupling, and g has the canonical dimension of Fig. 1a. The vanishing vortex tension allows monopoles to mass ([mass]). To describe materials with the relative electric break loose, while vortices themselves grow infinitely long permittivity ε and relative magnetic permeability µ, we incor- and loose and turn into unobservable Dirac strings, as shown √ porate the velocity of light v = 1/ εµ < 1 by defining the in Fig. 1b. Accordingly, the world-lines of monopoles be- Euclidean time lattice spacing as ` = `/v and by rescaling come infinitely long as well, which means that they Bose con- 0 all time derivatives, currents, and zero-components of gauge dense. Determining the condensation point is a dynamical fields by the factor 1/v. As a consequence, both gauge fields problem. One can say that the condensation point is set by p acquire a dispersion relation E = m2v4 + v2p2 with the topo- the moment when quantum corrections to the vortex tension logical mass [32] m = f g/2πv. The dimensionless parameter are large enough to turn it negative so that vortices become f = O(e) encodes the effective Coulomb interaction strength Dirac strings. The rest of this paper is devoted to answering in the material, g is the magnetic scale g = O(1/λ ), where λ this question and to deriving the nature of the ensuing new L L is the London penetration depth of the superconducting phase. state of matter. To analyze how the additional interactions can drive quan- To conclude here, we generalize the 2D consideration tum phase transitions taking the system out of the bosonic of [12, 23] onto 3D systems and predict that in three di- insulator, we integrate over the fictitious gauge fields to ob- mensions superconductors also acquire self-induced emergent tain a Euclidean action S for point charges and line vor- granularity in the vicinity of the SIT. Magnetic monopoles cv tices alone. As we show in Methods, this is proportional to play a crucial role in the formation and properties of this new the length of the charge world-lines and to the area of vor- superconducting state. tex world-surfaces, exactly as their configurational entropy. Charges and vortices can thus be assigned an effective action PHASE TRANSITIONS AND PHASE DIAGRAM (equivalent to a quantum “free energy” in this Euclidean field theory context)

Fcv = (sc − hc) N + (sv − hv) A , (3) Let us consider a granular system (irrespective to whether the granularity is the self-induced electronic granularity [13] or is where N and A are the length of word-lines in number of lat- of the structural origin, such as, e.g. granular diamond [30]), tice links and the area of world-surfaces in number of lat- characterized by the length scale ` playing the role of the gran- tice plaquettes, respectively and sc,v and hc,v denote the ac- ule size, and examine the various phases that can emerge. We tion and entropy contributions (per length and area) of charges focus on cubic lattices since, as in 2D, the paradigmatic sys- and vortices, respectively. The possible phases that are real- tem for monopoles and the transitions they induce is a Joseph- ized are determined by the relation between the values of the son junction array [31]. Of course, the universality class of Coulomb and magnetic scales and by materials parameters de- the transition can depend on the lattice details but the discus- termining whether the coefficients of N and A are positive or sion of such effects is beyond the scope of the present work. negative. For positive coefficients, long world-lines and large We thus formulate the action (1) on a cubic lattice of spac- wold-surfaces are suppressed. If either of the parenthesis be- ing ` and we add all possible local gauge invariant terms. In comes negative, then either a charge or a monopole conden- the presence of magnetic monopoles the tensor current mµν sate forms. In the case of charges, the proliferation of long- does not conserve any more, and the gauge invariance of the world lines is the geometric picture of Bose condensation first second kind of the tensor field bµν is effectively broken at put forth by Onsager [33] and elaborated by Feynman [34], the vortex endpoints. Longitudinal components of the ten- see [35] for a recent discussion. In the case of vortices, the sor gauge field bµν become the usual vector gauge fields for string between magnetic monopole endpoints becomes loose the magnetic monopoles and induce for the monopoles the and assumes the role of an unobservable Dirac string. In this same type of Coulomb interaction which experience electric case, the monopoles are characterized by long world-lines de- charges. However, this Coulomb interaction is subdominant scribing their Bose condensate phase, see Fig.1. The details to the linear tension created by vortices connecting monopole of this vortex transition have been discussed in [36, 37]. The and anti-monopole pair, and one can neglect it when deter- resulting phase diagram is determined by the value of the 4 √ parameter η = 2π(mv`)G/ µNµA encoding the strength of Taking the limit mv`  1, we find quantum fluctuations and the material properties of the sys- 1 P 2 p X − Fµν−2πtµν 8 f 2 x,µ,ν( ) tem [18] and tuning parameter, γ = ( f /`g) µN/µA, taking the exp (−S e.m.) = e . (5) system across superconductor-insulator transition (SIT). Here tµν G = O(G(mv`)), where G(mv`) is the diagonal element of the The monopole condensation for strong f renders the real elec- lattice kernel G(x − y) representing the inverse of the operator   tromagnetic field a compact variable, defined on the inter- `2 mv 2 − ∇2 µ µ ( ) , and N and A are the entropy per unit length val [−π, +π] and the electromagnetic response is given by of the world line and per unit area of the world surface, re- Polyakov’s compact QED action [40, 41]. This changes dras- spectively. The phase structure at T = 0 and the domains of tically the Coulomb interaction. To see that, let us take two different phases in the critical vicinity of the SIT are defined external probe charges ±qext and find the expectation value by the relations, see Methods for details: for the corresponding Wilson loop operator W(C), where C  is the closed loop in 4D Euclidean space-time (the factor ` is γ < 1 , charge Bose condensate , absorbed into the gauge field A to make it dimensionless) η < 1 →  µ γ > 1 , monopole Bose condensate , Z +π 1 P 2 1 X − Fµν−2πtµν P  8 f 2 x,µν( ) iqext C lµ Aµ γ < 1 , charge Bose condensate , hW(C)i = DAµ e e ,  η ZAµ,tµν −π  {tµν} η > 1 →  1 < γ < η , bosonic insulator , (4)  η (6)  γ > η , monopole Bose condensate . where lµ takes the value 1 on the links forming the Wilson loop C and 0 otherwise. When the loop C is restricted to and are shown in Fig. 2. The finite-temperature decay of the plane formed by the Euclidean time and one of the space the condensates, corresponding to the deconfinement transi- coordinates, hW(C)i measures the interaction energy between tions into the bosonic insulator, is described by the same ap- charges ±qext. A perimeter law indicates a short-range po- proach [38]. tential, while an area-law is tantamount to a linear interaction The effect of charge condensation is immediately revealed between them [41]. For Cooper pairs, qext = 1, see Meth- h i − by minimally coupling the current jµ = (1/2π)hµ to the elec- ods, W(C) = exp( σA) where A is the area of the surface S enclosed by the loop C. This yields a linear interaction be- tromagnetic gauge potential Aµ and integrating out all matter fields to obtain the electromagnetic response. The latter ac- tween probe Cooper pairs, which therefore can be viewed as µ confined by the elastic string with the string tension quires a photon mass term ∝ AµA so that the induced charge ! current jµ ∝ Aµ. This is nothing but London equation i.e. the 32 f 1 πG(0) major manifestation of superconductivity [16]. Not dwelling σ = √ exp − , (7) π εµ `2 8 f 2 on this standard derivation, but just stressing that our approach reproduces that charge condensate phase is a superconductor, where G(0)=0.155 is the value of the 4D lattice Coulomb po- we now focus on the new phase induced by the condensation tential at coinciding points. The monopole condensate, thus, of magnetic monopoles. At f /` < O(1), a superconducting generates a string binding together charges and preventing phase is thus realized, as observed in granular diamond [30]. charge transport in systems of a sufficient spatial size. A For `g/ f < O(1), however, there is a dual superinsulating magnetic monopole condensate is a 3D superinsulator, char- phase governed by the magnetic monopole condensate, which acterized by an infinite resistance at finite temperatures [12– is the main subject of this paper. 14, 17, 18]. The critical value of the effective Coulomb inter- To conclude this section, we would like to point out that action strength for the transition to the superinsulating phase the direct transition between superinsulator and supercon- is fcrit = O(`/λ). ductor for η < 1 shown in Fig. 2 matches perfectly the low-temperature, quantum phase structure of the QCD as a function of the density, with a transition between confined DISCUSSION hadronic matter and colour superconductivity [39].

Superinsulation has been observed in 2D, where magnetic monopoles are instantons rather than particles [13, 17, 18] ELECTROMAGNETIC RESPONSE AND THE ELECTRIC and the structure of the phase diagram is conclusively es- STRING TENSION tablished. The charge Berezinskii-Kosterlitz-Thouless (BKT) transition into the low-temperature confined superinsulating phase was measured. Reference [18] presented measurements To reveal the nature of the superinsulating phase, we exam- of the electromagnetic response of the superinsulating phase ine its electromagnetic response. To that end we again min- and of properties of electric strings. In particular, the observed imally couple the electric current jµ to the electromagnetic double kinks in the I-V characteristics indicate the predicted gauge field Aµ and compute its effective action, see Methods. electric Meissner effect characteristic to superinsulators. The 5

e DATA AVAILABILITY bosonic insulator: no condensates phase Topological insula Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

magnetic monopole Cooper pair condensate: Acknowledgements condensate:Superinsulator Superconductor Superconductor V.M.V. thanks Terra Quantum for support at the final stage Superinsulator of the work. M.C.D. thanks CERN, where she completed this work, for kind hospitality. tuning parameter, �

FIG. 2. Cooper pairs–quantum magnetic monopoles phase dia- √ APPENDIX gram at T=0. The parameter η = 2π(mv`)G/ µNµA encodes the strength of quantum fluctuations and the material properties of the system. The inter-phases lines are η=1/γ (γ<1, η> 1), γ=1 (η<1), Lattice BF term and η=γ (η>1, γ>1) .

To formulate the gauge-invariant lattice BF-term, we fol- low [12] and introduce the lattice BF operators experimental results exhibit a fair agreement with the theoreti- cal predictions. The existence of a bosonic insulator in various k ≡ S  d , materials and, in particular, in the same NbTiN films where µνρ µ µανρ α ˆ ˆ ˆ the kinked I-V curves were measured, was unambiguously es- kµνρ ≡ µναρdαS ρ , (8) tablished in [42]. That TiN films with smaller η < 1 exhibit the direct SIT, while NbTiN films endowed with η > 1 show where the SIT across the intermediate topological insulator phase, complies with the theoretical expectations. However, more f (x + `µˆ) − f (x) d f (x) ≡ , S f (x) ≡ f (x + `µˆ) , research on various systems is required to spot the exact loca- µ ` µ tion of the tricritical point. To establish the Cooper pair-based f (x) − f (x − `µˆ) dˆ f (x) ≡ , Sˆ f (x) ≡ f (x − `µˆ) , (9) nature of the bosonic topological insulator, shot noise mea- µ ` µ surements similar to those carried out in [43] are desirable. The signature of 3D superinsulation [44] has been detected in are the forward and backward lattice derivative and shift op- InO films [45, 46]. However, more studies are necessary for erators, respectively. Summation by parts on the lattice inter- drawing reliable conclusions that InO can be considered as a changes both the two derivatives (with a minus sign) and the material hosting a magnetic monopole condensate. Strongly two shift operators; gauge transformations are defined using type II superconductors with a fine, inherent or self-induced the forward lattice derivative. The two lattice BF operators granular electronic texture, are other most plausible candi- are interchanged (no minus sign) upon summation by parts on dates to house 3D superinsulators. Finally, another class of the lattice and are gauge invariant so that: candidates are layered materials. Vortex lines in such mate- rials can be regarded as stacks of pancake vortices [24]. If k d = k d = dˆ k = 0 , the layers are weakly coupled and vortices are ballistic, the µνρ ν µνρ ρ µ µνρ ˆ ˆ ˆ ˆ ˆ pancakes can split and form magnetic monopoles at the inner kµνρdρ = dµkµνρ = dνkµνρ = 0 . (10) layers’ intersections, albeit very anisotropic ones. These can then condense into a superinsulating phase. And satisfy the equations Finally, while the monopole condensate existence is   2 strongly supported by the observation of the superinsulating kˆµνρkρλω = − δµλδνω − δµωδνλ ∇ state and the corresponding experimental implications are re-     ˆ − ˆ ˆ − ˆ liably established by recent transport measurements [18], the + δµλdνdω δνλdµdω + δνωdµdλ δµωdνdλ ,  2  conclusive evidence for monopoles should come from their kˆµνρkρνω = kµνρkˆρνω = 2 δµω∇ − dµdˆω , (11) direct observations. One of the ways to implement such an ob- servation may be extending the SQUID-on-tip device method 2 ˆ of [4] to lower temperatures. where ∇ = dµdµ is the lattice Laplacian. We use the notation ∆µ and ∆ˆ µ for the forward and backwards finite difference operators. 6

Phases of monopoles are gapped excitations, suppressed by their large action. If the coefficients, instead are negative, the entropy dominates and To find the topological action for monopoles, we start from large configurations are favoured in the “free energy” (effec- eq. (2) and integrate out fictitious gauge fields aµ and bµν tive action). The phase in which long world-lines of Cooper 2 2 pairs dominate the Euclidean partition function is a charge X f δµν g δµαδνβ − δµβδνα S = q q + m m Bose condensate, as discussed originally by Onsager [33] and top `2 µ mv 2 − ∇2 ν µν mv 2 − ∇2 αβ x 2 ( ) 8 ( ) Feynmann [34] (for a recent discussion see [35]). This phase 2 π(mv) kµαβ is the Bose condensate of magnetic monopoles. For vortices, +i q m(12), ` µ ∇2 (mv)2 − ∇2
αβ proliferation of large world-surfaces means that the strings binding monopoles and antimonopoles into neutral pairs be- The last term represents the Aharonov-Bohm phases of come loose. We will show below that in this case the long real λ charged particles around vortices of width L. On scales much monopole world-lines dominate the electromagnetic response. λ mv 2∇2 larger than L, where the denominator reduces to ( ) , The combined energy-entropy balance equations are best − this term becomes i2π integer, as can be easily recognized viewed as defining the interior of an ellipse on a 2D integer by expressing qµ = (1/2)`kµαβyαβ. This reflects the absence lattice of electric and magnetic quantum numbers, of Aharonov-Bohm phases between charges ne and magnetic fluxes 2π/ne. Accordingly, we shall henceforth neglect this Q2 M2 + < 1 , (15) term. 2 2 rQ rM The important consequence of the topological interactions is that they induce self-energies in form of the mass of Cooper where the semiaxes are given by pairs and tension for vortices between magnetic monopoles. `g 1 `g rµ 1 These self-energies are encoded in the short-range kernels in r2 = = N , Q f η f µ η the action (12), which we approximate by a constant. World- Q A r lines and world-surfaces are thus assigned “energies” (for- 2 f 1 f µA 1 rM = = , (16) mally Euclidean actions in the present statistical field theory `g ηM `g µN η setting and thus dimensionless in our units) proportional to with their length N and area A (measured in numbers of links and plaquettes), √ √ η = ηQηM = 2π(mv`)G/ µNµA . (17) f 2 S N = 2π(mv`)G Q N , Q M g` Of course, configurations with , 0 and , 0 must be excluded since the two types of excitations are different, g` 2 S A = 2π(mv`)G M A , (13) only pairs {0, M} or {Q, 0} have to be considered. The phase f diagram is found by establishing which integer charges lie where G = O(G(mv`)), with G(mv`) the diagonal element of within the ellipse when the semi-axes are varied. This yields the lattice kernel G(x − y) representing the inverse of the op-   Eq. (4) in the main text. erator `2 (mv)2 − ∇2 , and Q and M are the integer quantum numbers carried by the two kinds of topological defects. How- ever, also the entropy of link strings and plaquette surfaces is Electromagnetic response in the magnetic monopole condensate proportional to their length and area [47], µN N and µA A. Both coefficients µ are non-universal: µN ' ln(7) since at each step the non-backtracking string can choose among 7 possible di- To establish the electromagnetic response of the monopole rections on how to continue, while µA does not have such a condensate we add the minimal coupling of the charge current simple interpretation but can be estimated numerically. This jµ to the electromagnetic field, gives for both types of topological defects a “free energy” pro- X X 1 portional to their dimension and with coefficients that can be S → S − i `4A j = S − i `4 A k b , (18) µ µ 4π µ µαβ αβ positive or negative depending on the parameters of the the- x x ory. The total free energy is and we compute its effective action by integrating over the " ! ! # F f 1 g` 1 fictitious gauge fields a and b . This requires no new com- = Q2 − N + M2 − A , µ µν 2π(mv`)G g` ηQ f ηM putation since, by a summation by parts, the above coupling amounts only to a shift where we have defined 2π(mv`)G 2π(mv`)G 1 2 ˆ ηQ = , ηM = . (14) mµν → mµν − ` kµναAα , (19) µN µA 2π

If the coefficients are positive, the self-energy dominates and in (12). Setting qµ = 0 for the phase with gapped Cooper large string/surface configurations are suppressed in the par- pairs gives Eq. (5) in the main text. tition function. In this regime Cooper pairs and/or vortices 7

Computation of the string tension By shifting the gauge field χµ by −qextηµ and introducing M2 = (π2/2 f 2)z, we can rewrite this as

The starting point is equation (6) in the main text. For large Z 2 1 2 f P 1 0 2 2 − x,µ,ν gµν +M (1−cos(χµ)) values of the coupling f , the action is peaked around the val- hW(C)i = Dχµ e π2 4 , (26) Zχ ues Fµν = 2πtµν, allowing for the saddle-point approximation µ to compute the Wilson loop. Using the lattice Stoke’s theo-   0 − rem, one rewrites Eq. (6) as where gµν = gµν χµ qextηµ . For large f , this integral is dominated by the classical solution to the equation of motion Z +π X 1 P ˜ 2 q P 1 − x(Fµν−2πmµν) i ext S (F˜ −2πm ) hW(C)i = DA e 8 f 2 e 2 S µν µν µν , µ ∆ˆ gcl = −2πq ∆ˆ S + M2sinχcl . (27) ZAµ,mµν −π µ µν ext µ µν ν {mµν} (20) Let us assume that the Wilson loop lies in the (0-3) plane where the quantities S µν are unit surface elements perpendic- formed by the Euclidean time direction 0 and the z axis. In ular (in 4D) to the plaquettes forming the surface S encircled this case, there are non-trivial solutions only for the 1- and by the loop C and vanish on all other plaquettes. We have cl 2-components of the gauge field, while χ3 = 0. With the also multiplied the Wilson loop operator by 1 in the form Ansatz χcl = χcl(x ), χcl = χcl(x ), we are left with two one- − P 1 1 2 2 2 1 exp( iπqext x S µνmµν). Following Polyakov [41], we decom- dimensional equations in the region far from the boundaries pose mµν into transverse and longitudinal components, of the Wilson surface S , T L mµν = m µν + m µν , cl 2 cl ∆ˆ 1∆1χ = −2πqext∆ˆ 1S 12 + M sinχ , mT = `kˆ n + `kˆ ξ , 2 2 µν µνα α µνα α ˆ cl ˆ 2 cl L ∆2∆2χ1 = −2πqext∆2S 21 + M sinχ1 , (28) m µν = ∆µλν − ∆νλµ , (21) Following [41], we solve these equations in the continuum where {nµ} are integers and we adopt the gauge choice ∆µλµ = 2 limit, 0, so that ∇ λµ = ∆ˆ ν∆νλµ = mµ, with mµ ∈ Z describe the world-lines of the magnetic monopoles on the lattice. The set cl 0 2 cl ∂1∂1χ2 = 2πqextδ (x1) + M sinχ2 , of 6 integers {mµν} has thus been traded for 3 integers {nµ} and cl 0 2 cl ∂2∂2χ1 = 2πqextδ (x2) + M sinχ1 . (29) 3 integers {mµ} representing the magnetic monopoles. The for- mer are then used to shift the integration domain for the gauge For qext = 1 (corresponding to Cooper pairs in our case), the field Aµ to [−∞, +∞]. The real variables {ξµ} can then also cl classical solutions with the boundary conditions χ1,2 → 0 for be absorbed into the gauge field. The integral over the now |x1,2| → ∞ are non-compact gauge field Aµ gives the Gaussian fluctuations cl −M|x2| around the saddle points mµ. Gaussian fluctuations contribute χ1 = sign(x2) 4 arctan e , | | cl −M|x1| the usual Coulomb potential 1/ x in 3D. We shall henceforth χ2 = sign(x1) 4 arctan e . (30) focus only on the magnetic monopoles. Inserting this back in (26) we get formula (7) in the main text. 2 X π P 1 P 1 1 − x,µ mµ mµ i2πq m ∆ˆ S hW(C)i = e 2 f 2 −∇2 e ext S µ −∇2 ν νµ . Zmµ {mµ} (22) Following [48] we introduce a dual gauge field χ with field [1] Goddard, P. & Olive, D. I. Magnetic monopoles in gauge field µ theories. Rep. Prog. Phys. 41, 1357–1437 (1978). − strength gµν = ∆µχν ∆νχµ and we rewrite (22) as [2] Milton, K. A. Theoretical and experimental status of magnetic Z 2 N monopoles. Rep. Prog. Phys. 69 1637-1712 (2006). 1 f P 2 X z X X P − x,µ,ν gµν i x,µ mµ(χµ+qextηµ) hW(C)i = Dχµ e 2π2 e [3] Qi, X. L.,, Li, R., Zhang, J. & Zhang, S.-C. Inducing a magnetic Zmµ,χµ N! 1 n N x1,...,xN mµ,...,mµ=±1 monopole with topological surface states. Science 323, 1184– (23) 1187 (2009). 2 [4] Uri, A. et al. Nanoscale imaging of equilibrium quantum Hall where the angle ηµ = 2π∆ˆ νS νµ/(−∇ ) represents a dipole sheet edge currents and of the magnetic monopole response in on the Wilson surface S and the monopole fugacity z is deter- graphene. Nature Physics 16, 164–170 (2020). mined by the self-interaction as [5] Dirac, P.A. M. Quantised singularities in the electromagnetic 2 − π G(0) field. Proc. R. Soc A 133, 60–72 (1931). z = e 2 f 2 , (24) [6] ’t Hooft, G. Magnetic Monopoles in Unified Gauge Theories. Nucl. Phys. B 79, 267–284 (1974). with G(0) being the inverse of the Laplacian at coincid- [7] Polyakov, A. M. Particle spectrum in quantum field theory. ing arguments. We also used the dilute gas approximation, JETP Lett. 20, 194-195 (1974). valid at large f , in which one takes into account only single [8] Nambu, Y. Strings, monopoles and gauge fields. Phys. Rev. D10 monopoles mµ = ±1. The sum can now be explicitly per- 4262 (1974). formed [48], with the result, [9] Mandelstam, S. Vorticesand quark confinement in non-Abelian gauge theories. Phys. Rep. 23C 245-249 (1976). 2 Z f P 2 4π2 1 − 2 x,µ,ν gµν+ 2 z(1−cos(χµ+qextηµ)) [10] ’t Hooft, G. On the phase transition towards permanent quark hW(C)i = Dχµ e 2π f , (25) Zχµ confinement. Nucl. Phys. B138 1346-1349 (1978). 8

[11] Ya. M. Blanter, R. Fazio, and G. Schon.¨ Duality in Josephson Phys. B 437, 695–722 (1995). Junction Arrays. Nuclear Physics B (Proc. Suppl.) 58, 79 – 90 [29] Vishwanath, A. & Senthil, T. Physics of Three-Dimensional (1997). Bosonic Topological Insulators: Surface-Deconfined Criticality [12] Diamantini, M. C., Sodano, P. & Trugenberger, C. A. Gauge and Quantized Magnetoelectric Effect. Phys. Rev. X 3, 011016 theories of Josephson junction arrays. Nuclear Physics B474, (2013). 641 – 677 (1996). [30] Zhang, G. et. al. Metal-bosonic insulator-superconductor tran- [13] Vinokur, V.M. et al. Superinsulator and quantum synchroniza- sition in boron-doped granular diamond. Phys. Rev. Lett 110, tion. Nature 452, 613 – 615 (2008). 077001 (2013). [14] Diamantini, M. C., Trugenberger, C. A. & Vinokur, V.M. Con- [31] Fazio, R. and van der Zant, H. Quantum phase transitions and finement and asymptotic freedom with Cooper pairs. Comm. vortex dynamics in superconducting networks. Physics Reports Phys. 1, 77 (2018). 355, 235 – 334 (2001). [15] Trugenberger, C., Diamantini, M. C., Nogueira, F. S., Poccia, N. [32] Allen, T., Bowick, M. & Lahiri, A. Topological Mass Genera- & Vinokur, V.M. Magnetic Monopoles and Superinsulation in tion in 3+1 dimensions. Mod. Phys. Lett. A 6, 559–571 (1991). Josephson Junction Arrays. Quantum Reports 2, 388 – 399 [33] Onsager, L. Statistical hydrodynamics Nuovo Cimento Supp. 6, (2020). 279–287 (1949). [16] Tinkham, M. Introduction to superconductivity. McGraw-Hill, [34] Feynman, R. Statistical Mechanics Benjamin, Reading (1972). Inc. 1996. [35] Schakel, A. M. J. Percolation, Bose-Einstein condensation and [17] Mironov, A. Yu. et al.Charge Berezinskii-Kosterlitz-Thouless string proliferation. Phys. Rev. E 63, 026115 (2001). transition in superconducting NbTiN films. Scientific Reports [36] Kleinert, H. & Chervyakov, A. Evidence for negative stiffness 8 4082 (2018). of QCD flux tubes in the large-N limit of SU(N). Physics Letters [18] Diamantini, M. C., Gammaitoni, L., Strunk, C., Postolova, S. V., B 381, 286–290 (1996). Mironov, A Yu., Trugenberger, C. A. & Vinokur, V.M. Direct [37] Diamantini, M. C. & Trugenberger, C. A. Geometric aspects of probe of the interior of an electric pion in a Cooper pair su- confining strings. Nucl. Phys. B 531, 151–167 (1998). perinsulator. Communications Physics 3, 142 (2020). [38] Diamantini, M. C. et al. Bosonic topological insulator interme- [19] Greensite, J. An introduction to the confinement problem. diate state in the superconductor-insulator transition. Physics Springer-Verlag (Berlin) 2011. Letters A 384, 126570 (2020). [20] Kowal, D. & Ovadyahu, Z. Disorder induced granularity in an [39] Schmidt, C. & Sharma, S. The phase structure of QCD. Jour. amorphous superconductor. Solid St. Comm. 90,783 – 786 Phys. G: Nucl. Part. Phys. 44 104002 (2017). (1994). [40] Polyakov, A. Compact gauge fields and the infrared catastrophe. [21] Fistul, M. V., Vinokur, V.M. &Baturina, T.I.˙ Collective Cooper- Physics Letters B 59, 82–84 (1975). Pair Transport in the Insulating State of Josephson-Junction Ar- [41] Polyakov, A. M. Gauge Fields and Strings, Harwood Academic rays. Phys. Rev. Lett. 100, 086805 (2008). Publisher, Chur (Switzerland) (1987). [22] Baturina, T. I. & Vinokur, V.M. Superinsulator–superconductor [42] Diamantini, M. C. et al. Bosonic topological insulator interme- duality in two dimensions. Annals of Physics, 331, 236 – 257 diate state in the superconductor-insulator transition. Physics (2013). Letters A 384, 126570 (2020). [23] Diamantini, M. C., Trugenberger, C. A. & Vinokur, V.M. Topo- [43] Zhou, P. et al, Electron pairing in the pseudogap state revealed logical gauge theory of the superconductor-insulator transition. by shot noise in copper oxide junctions. Nature 572, 493–496 In Topological Phase Transitions and New Developments, pp (2019). 197 – 221, World Scientific, 2019. [44] Diamantini, M. C., Gammaitoni, L., Trugenberger, C. A. & Vi- [24] Blatter, G., Feigel’man, M. V., Geshkenbein, V.B., Larkin, A. I. nokur, V.M. Voger-Fulcher-Tamman criticality of 3D superin- & Vinokur, V.M. Vortices in high-temperature supercondcu- sulators, Scientific Reports 8, 15718 (2018). tors. Rev. Mod. Phys. 66, 1125–1388 (1994). [45] Sambandamurthy, G., Engel, L. M., Johansson, A., Peled, E. [25] Wilczek, F., Disassembling Anyons. Phys. Rev. Lett. 69 132- & Shahar, D. Experimental evidence for a collective insulating 135 (1992). state in two-dimensional superconductors.Phys. Rev. Lett. 94, [26] Birmingham, D., Blau, M., Rakowski, M. & Thompson, G. 017003 (2005). Topological field theory Phys. Rep. 209, 129–340 (1991). [46] Ovadia, M. et al. Evidence for a finite-temperature insulator. [27] Wen, X.-G. Topological Order: From Long-Range Sci. Rep. 5, 13503 (2015). Entangled Quantum Matter to a Unified Origin of [47] Nelson, D., Piran, T., & Weinberg, S. eds. Statistical Mechan- Light and Electrons. ISRN 2013, 198710 (2013). ics of Membranes and Surfaces, World Scientific, Singapors https://doi.org/10.1155/2013/198710 (2004). [28] Bergeron, M., Semenoff G. W. & Szabo, R. Canonical BF-type [48] Orland, P. Instantons and disorder in antisymmetric tensor topological field theory and fractional statistics of strings. Nucl. gauge fields. Nucl. Phys. B 205, 107-118 (1982).