betaFeedback

WorksPhysics and Astronomy

Ohm's law, Joule heat, and Planckian dissipation

Abstract

Electric current generation and its dissipation are important physical processes. It ranges from the one follows the Ohm's law to superconductivity. Recently, it has been shown that the gradient of the chemical potential force arises from the time-component of the Berry connection from many-electron wave functions, and we consider its importance for the electric current conduction in this work. We first show that it rectifies the odd explanation in Joule heating by electric current in a metallic wire: Poynting's theorem explains that the energy for the Joule heating enters from the outside of the wire as radiation. We show that this energy is supplied by the chemical potential gradient generated by the battery connection. Next, we consider the discharging of a capacitor problem where the capacitor plays a role of a battery; and the tunneling supercurrent through the Josephson junction problem, where the original derivation did not include the capacitor contribution. Lastly, we argue that the gauge fluctuation of the time-component of the Berry connection included in the chemical potential gradient force might explain the Planckian dissipation observed in high transition temperature cuprate superconductors. The present work suggests the rethinking of the gauge invariance in Maxwell's equations.

Introduction

One of old and established physical laws for the electric current is Ohm’s law discovered by Ohm [1] (unpublished materials indicate it was known before Ohm by Cavendish [2]): The electric current II generated by connecting a battery of voltage VV to a metallic wire with resistance RR is given by

I=VR(1)I = \frac{V}{R} \tag*{(1)}

The battery originally used by Ohm was an electric battery, where the electromotive force was generated by chemical reactions.

Joule found that the heat generated by a metallic wire of resistance RR with current II is given by

RI2=IV(2)RI^{2} = IV \tag*{(2)}

where the Ohm’s law is used [3]. It indicates the heat is equivalent to the work done by the battery, establishing that the heat is a kind of energy.

Surprisingly, this old, thought to be established, dissipation phenomenon has loose ends that should be tied: The energy flow explained by the Poynting theorem for the above Joule heat generation is known to be rather odd [4, 5].

Let us first explain the Poynting theorem. The conservation of the energy in systems composed of electromagnetic field and electric current was formulated by Poynting and called, ‘Poynting’s theorem’ [6]. It states that the rate of the energy consumed by the current in the volume V\mathcal{V} is given by ∫Vd3r j⋅E\int_{\mathcal{V}} d^{3}r\,\mathbf{j}\cdot\mathbf{E}, where j\mathbf{j} and E\mathbf{E} are the current density and electric field, respectively; and the rate of the energy leaving V\mathcal{V} as radiation is given by ∫SdS⋅(E×H)\int_{S}d\mathbf{S}\cdot(\mathbf{E}\times\mathbf{H}), where B\mathbf{B} is the magnetic field, H\mathbf{H} is related to B\mathbf{B} by H=μ0−1B\mathbf{H}=\mu_{0}^{-1}\mathbf{B} (μ0\mu_{0} is the vacuum permeability) in a vacuum, and SS is the boundary surface of the volume V\mathcal{V}. In short, the energy conservation is given by

∫Vd3r j⋅E+∫SdS⋅(E×H)=−∂∂t∫Vd3r u(3)\int_{\mathcal{V}} d^{3}r\,\mathbf{j}\cdot\mathbf{E}+\int_{S}d\mathbf{S}\cdot(\mathbf{E}\times\mathbf{H})=-\frac{\partial}{\partial t}\int_{\mathcal{V}}d^{3}r\,u \tag*{(3)}

where uu is the energy density of the electromagnetic field. The first term looks similar to Eq. (2); thus, it is often said that it expresses the Joule heat.

The Ohm’s law was explained by Drude [7], and later by Lorentz [8, 9] and Bohr [10], assuming that the battery connected to the wire generates the electric field EDrudeE_{\mathrm{Drude}} inside it. The law is expressed in a local field relation,

j=σEDrude(4)\mathbf{j}=\sigma E_{\mathrm{Drude}} \tag*{(4)}

where σ\sigma is the conductivity of the wire. Then, j⋅E\mathbf{j}\cdot\mathbf{E} in the first term in Eq. (3) is given by j⋅EDrude=σ−1j2>0\mathbf{j}\cdot\mathbf{E}_{\mathrm{Drude}}=\sigma^{-1}j^{2}>0; here, the plus sign indicates that it is the energy consumed.

The oddness of the energy flow is that the radiation enters into the wire from outside as expressed by the Poynting vector E×H\mathbf{E}\times\mathbf{H} in Eq. (3), and supplies the energy for the Joule heat consumed in the wire. However, it is sensible to consider that the energy for the Joule heat is supplied by the battery through the current flows in the wire.

This problem was taken up by the present author, recently [11]. A sensible explanation is obtained by identifying that the gradient of the chemical potential force is the one accelerates electrons instead of the electric field. The chemical potential was originally introduced by Gibbs (he called it, the intrinsic potential) in the context of thermodynamics [12]. During the time when Drude developed his theory, the chemical potential was unknown; however, we now know the chemical potential exists, and the electromotive force generated is due to the chemical potential difference originates from chemical reactions in the battery. The chemical potential appears as the Fermi energy in solid state physics, and its gradient is known to produce current. It is also used in the Landauer-Büttiker theory [13, 14, 15], where the chemical potential appears as the Fermi energy of the distribution function of the free-electron theory of conduction electrons.

It has been shown recently that this chemical potential arises as a quantum many-body effect described by the Berry connection [11]. This indicates the appearance of the Berry connection and also of the chemical potential are general quantities of quantum many-body systems [16, 17] that include ℏ\hbar (ℏ\hbar is the reduced Planck constant).

Recently, the Planckian dissipation observed in strange metals such as high transition temperature cuprate superconductors is a focus of attention in condensed matter physics [18, 19, 20]. It is characterized by the relaxation time with ℏ\hbar, given by τℏ=ℏkBT\tau_{\hbar} = \frac{\hbar}{k_{\mathrm{B}}T} (kBk_{\mathrm{B}} is the Boltzmann constant, and TT is temperature). The elucidation of the origin of it is considered to be the key to understand high transition temperature cuprate superconductivity. One of the purposes of the present work is to show that this Planckian dissipation may be related to the fluctuation of the Berry connection that produces the chemical potential. If this is the case, the loose ends of the Joule heating problem and the Planckian dissipation in the cuprate superconductivity have the common origin.

The Berry connection from many-electron wave functions was originally put forward by the present author to develop a new superconductivity theory [16, 17] that encompasses the standard theory based on the BCS one [21]. The initial motivation for the development of it is to elucidate the mechanism of the cuprate superconductivity; however, during such an effort, it became apparent that experimental facts of superconductivity indicate the presence of several serious loose ends to be tied in the standard theory (we will discuss them in Section 4). The new theory was made to tie them with keeping major successful results of the BCS theory intact. This is achieved by introducing new Bogoliubov type excitations; they preserve the particle number unlike the original ones, and the Nambu-Goldstone like mode arises form the Berry connection with keeping the particle number fixed (A short account of this theory and its relation to the BCS theory is given in A).

In the new theory, the Berry connection produces persistent current if it is equipped with nontrivial topological quantum numbers. If the fluctuation of the topological quantum numbers for the Berry connection (we will call it, the ‘gauge fluctuation’) occurs, the current becomes resistive. Such a resistive current is expected to exist in the vicinity of a superconducting phase of the phase diagram, and may exhibit the Planckian dissipation. We will consider this possibility in the present work.

Before dealing with the Planckian dissipation problem, we will first revisit the Joule heating problem and show how the standard textbook explanation is rectified to a sensible explanation in which the energy is supplied from the battery through the current. We also consider the discharging of a capacitor; then, the supercurrent flow. through the Josephson junction including the capacitor contribution. It is noteworthy that although the Josephson junction inherently has a capacitor contribution due to its structure, it was not included in the original Josephson’s work [22] or the textbook explanation [23]. The derivation including the capacitance contribution indicates that the electrons transfer singly across the junction, in contrast to the Josephson’s derivation that assumes the pairwise transfer [22]. Note that this does not mean the absence of the stabilization by the electron-pairing; the experimentally observed flux quantum is still obtained.

Actually, the existence of single electrons in superconductors has been speculated for a long time to explain the Knight shift experiments [24]. It is also noteworthy that a recent experiment indicates the absence of the dissipative quantum phase transition in a Josephson junction system predicted by the standard theory [25]; and also the so-called ‘quasiparticle poisoning problem’ indicates the existence of a large amount of excited single electrons in Josephson junction systems for qubits, obtaining the observed ratio of their number to the Cooper pair number 10−9∼10−510^{-9} \sim10^{-5} in disagreement with the standard theory ratio 10−5210^{-52} [26, 27].

The organization of the present work is as follows: First, we briefly summarize the Berry connection from many-electron wave functions, and explain the appearance of the gradient of the chemical potential force from it. Next, we revisit the plasma oscillation and screening in the electron gas model including the gradient of the chemical potential force; it is indicated the force balance exists for the force from the static electric field and that from the chemical potential gradient. A similar force balance is assumed in the Ohm’s law situation, later. Then, we will proceed to put forward a new criterion to differentiate the normal and superconducting states based on the stability of the velocity field with the Berry connection equipped with non-trivial topological quantum numbers.

Next, we consider the energy flow for electric conduction in a metallic wire connected to a battery; we argue that the battery connection generates the gradient of the chemical potential inside the wire, and this is the force accelerates conduction electrons; when the steady current flow is established, the force balance exists between the gradient of the chemical potential force, and that from the induced electric field. Then, we consider the energy flow during the discharging of a capacitor through a resistive wire; in this case, the capacitor acts as the battery for the current generation. Next, we discuss the Josephson effect in superconductors with including the capacitor contribution. Then, the Planckian dissipation is examined by considering the gauge fluctuation of the Berry connection. Lastly, we conclude the present work by discussing its implications in the interpretation of the Maxwell’s equations.

Forces on electrons in an electromagnetic field including the Berry connection from many-electron wave functions

Let us consider the Berry connection from many-body wave functions defined by

AΨMB(r,t)=Re⁡{∫dσ1dx2⋯dxN Ψ∗(r,σ1,…,xN,t)(−iℏ∇)Ψ(r,σ1,…,xN,t)}ℏN−1ρ(r,t)(5)\mathbf{A}^{\mathrm{MB}}_{\Psi}(\mathbf{r},t)=\frac{\operatorname{Re}\left\{\int d\sigma_{1}dx_{2}\cdots dx_{N}\,\Psi^{*}(\mathbf{r},\sigma_{1},\ldots,x_{N},t)(-i\hbar\nabla)\Psi(\mathbf{r},\sigma_{1},\ldots,x_{N},t)\right\}}{\hbar N^{-1}\rho(\mathbf{r},t)} \tag*{(5)}

Here, ‘Re’ denotes the real part, Ψ\Psi is the total wave function, xi=(ri,σi)\mathbf{x}_{i}=(\mathbf{r}_{i},\sigma_{i}) collectively stands for the coordinate ri\mathbf{r}_{i} and the spin σi\sigma_{i} of the iith electron, however, x1\mathbf{x}_{1} is expressed as x1=(r,σ1)\mathbf{x}_{1}=(\mathbf{r},\sigma_{1}); NN is the total number of electrons; −iℏ∇-i\hbar\nabla is the Schrödinger’s momentum operator for the coordinate vector r\mathbf{r}, and ρ\rho is the number density calculated from Ψ\Psi [11, 17].

From the original wave function Ψ\Psi, we can construct the following currentless wave function using the Berry connection AΨMB\mathbf{A}^{\mathrm{MB}}_{\Psi}

Ψ0(x1,…,xN,t)=Ψ(x1,…,xN,t)exp⁡(−i∑j=1N∫0rjAΨMB(r′,t)⋅dr′)(6)\Psi_{0}(x_{1},\ldots,x_{N},t)=\Psi(x_{1},\ldots,x_{N},t)\exp\left(-i\sum_{j=1}^{N}\int_{0}^{\mathbf{r}_{j}}\mathbf{A}^{\mathrm{MB}}_{\Psi}(\mathbf{r}',t)\cdot d\mathbf{r}'\right) \tag*{(6)}

The current density calculated by Ψ0\Psi_{0} is zero due to the cancellation of the current density from Ψ\Psi and that from exp⁡(−i∑j=1N∫0rjAΨMB(r′,t)⋅dr′)\exp\left(-i\sum_{j=1}^{N}\int_{0}^{\mathbf{r}_{j}}\mathbf{A}^{\mathrm{MB}}_{\Psi}(\mathbf{r}',t)\cdot d\mathbf{r}'\right). Conversely, Ψ\Psi is expressed using Ψ0\Psi_{0} as

Ψ(x1,…,xN,t)=exp⁡(i∑j=1N∫0rjAΨMB(r′,t)⋅dr′)Ψ0(x1,…,xN,t)(7)\Psi(x_{1},\ldots,x_{N},t)=\exp\left(i\sum_{j=1}^{N}\int_{0}^{\mathbf{r}_{j}}\mathbf{A}^{\mathrm{MB}}_{\Psi}(\mathbf{r}',t)\cdot d\mathbf{r}'\right)\Psi_{0}(x_{1},\ldots,x_{N},t) \tag*{(7)}

where the electric current generation is entirely attributed to the factor

exp⁡(i∑j∫0rjAΨMB(r′,t)⋅dr′)(8)\exp\left(i\sum_{j}\int_{0}^{\mathbf{r}_{j}}\mathbf{A}^{\mathrm{MB}}_{\Psi}(\mathbf{r}',t)\cdot d\mathbf{r}'\right) \tag*{(8)}

This acts as the collective motion of electrons that generates electric current.

The velocity field v\mathbf{v} for the electrons is calculated using Ψ\Psi with including the electromagnetic vector potential A\mathbf{A} as

v=emeA+ℏmeAΨMB(9)\mathbf{v}=\frac{e}{m_{e}}\mathbf{A}+\frac{\hbar}{m_{e}}\mathbf{A}^{\mathrm{MB}}_{\Psi} \tag*{(9)}

where −e-e is the electron charge and mem_e is the electron free mass. Up to this point, no approximation is made.

Now, we introduce angular variable χ\chi with period 2π2\pi and assume that it is related to AΨMB\mathbf{A}_{\Psi}^{\mathrm{MB}} in the following manner

χ(r,t)=−2∫0rAΨMB(r′,t)⋅dr′(10)\chi(\mathbf{r},t)=-2\int_{0}^{\mathbf{r}}\mathbf{A}_{\Psi}^{\mathrm{MB}}(\mathbf{r}',t)\cdot\mathrm{d}\mathbf{r}' \tag*{(10)}

The reason for the introduction of this χ\chi will be explained, later (see also A, around Eq. (84)).

Then, Ψ\Psi is expressed as

Ψ(x1,⋯ ,xN,t)=exp⁡(−i2∑j=1Nχ(rj,t))Ψ0(x1,⋯ ,xN,t)(11)\Psi(x_1,\cdots,x_N,t)=\exp\left(-\frac{i}{2}\sum_{j=1}^{N}\chi(\mathbf{r}_j,t)\right)\Psi_0(x_1,\cdots,x_N,t) \tag*{(11)}

and v\mathbf{v} becomes

v=eme(A−ℏ2e∇χ)(12)\mathbf{v}=\frac{e}{m_e}\left(\mathbf{A}-\frac{\hbar}{2e}\nabla\chi\right) \tag*{(12)}

Next, we consider the sum, A−ℏ2e∇χ\mathbf{A}-\frac{\hbar}{2e}\nabla\chi, in v\mathbf{v}. It is gauge invariant due to the fact that the gauge ambiguity of A\mathbf{A} is compensated by the ambiguity of ∇χ\nabla\chi; in other words, the value of A−ℏ2e∇χ\mathbf{A}-\frac{\hbar}{2e}\nabla\chi is the same irrespective of the gauge choice made for A\mathbf{A} since during the evaluation process of ∇χ\nabla\chi, this freedom is absorbed in the evaluated ∇χ\nabla\chi to give the same value for A−ℏ2e∇χ\mathbf{A}-\frac{\hbar}{2e}\nabla\chi (this point is shown in the calculations performed in [28]).

The gauge invariance of A−ℏ2e∇χ\mathbf{A}-\frac{\hbar}{2e}\nabla\chi implies its time-component partner, φ+ℏ2e∂tχ\varphi+\frac{\hbar}{2e}\partial_t\chi, is also gauge invariant. Note that the electromagnetic four-vector gauge potential (A,φ)(\mathbf{A},\varphi) becomes (A−ℏ2e∇χ,φ+ℏ2e∂tχ)(\mathbf{A}-\frac{\hbar}{2e}\nabla\chi,\varphi+\frac{\hbar}{2e}\partial_t\chi) with including the Berry connection. Taking into account the fact that the chemical potential μ\mu appears in a similar manner as the scalar potential φ\varphi in the electron Hamiltonian, and that the chemical potential is gauge invariant (however, φ\varphi is not), we obtain the following relation

μ=e(φ+ℏ2e∂tχ)(13)\mu=e\left(\varphi+\frac{\hbar}{2e}\partial_t\chi\right) \tag*{(13)}

This is the chemical potential we use in this work (this relation is also obtained as Eq. (106) in A from a different reasoning). Actually, a similar relation is obtained by Anderson for the case of superfluid helium [29].

Let us calculate the force acting on electrons by medvdtm_e\frac{\mathrm{d}\mathbf{v}}{\mathrm{d}t},

medvdt=me[∂tv+12∇v2−v×(∇×v)]m_e\frac{\mathrm{d}\mathbf{v}}{\mathrm{d}t}=m_e\left[\partial_t\mathbf{v}+\frac{1}{2}\nabla v^2-\mathbf{v}\times(\nabla\times\mathbf{v})\right]
=e∂tA−ℏ2∂t(∇χ)+me2∇v2−ev×B+ℏ2v×∇×(∇χ)=−eE−∇μ+me2∇v2−ev×B+ℏ2v×∇×(∇χ)+ℏ2(∇∂t−∂t∇)χ(14)\begin{aligned} &= e\partial_t\mathbf{A}-\frac{\hbar}{2}\partial_t(\nabla\chi)+\frac{m_e}{2}\nabla v^2-e\mathbf{v}\times\mathbf{B}+\frac{\hbar}{2}\mathbf{v}\times\nabla\times(\nabla\chi)\\ &=-e\mathbf{E}-\nabla\mu+\frac{m_e}{2}\nabla v^2-e\mathbf{v}\times\mathbf{B}\\ &\quad+\frac{\hbar}{2}\mathbf{v}\times\nabla\times(\nabla\chi)+\frac{\hbar}{2}(\nabla\partial_t-\partial_t\nabla)\chi \tag*{(14)} \end{aligned}

where Eqs. (12) and (13), and the following relations, B=∇×A\mathbf{B}=\nabla\times\mathbf{A}, E=−∂tA−∇φ\mathbf{E}=-\partial_t\mathbf{A}-\nabla\varphi are used. The forces in Eq. (14) include the standard Lorenz force −eE−ev×B-e\mathbf{E}-e\mathbf{v}\times\mathbf{B}, and additional forces. The important one in this work is the gradient of the chemical potential force, −∇μ-\nabla\mu.

At this point, it is worth mentioning the argument given by Maxwell in his book on electromagnetism [30]: The electromotive intensity at any point is the resultant force on a unit of positive electricity placed at that point. It may arise from the following three actions

  1. Electrostatic action

  2. Electromagnetic induction

  3. Thermoelectric or electrochemical action

The gradient of the chemical potential force belong to the third category. The first two are included in the standard textbook, but the third one is often not mentioned. There are terms in Eq. (14) that are not categorized above. They are important in superconductivity as shown in our previous work [11].

In superconductors, v\mathbf{v} in Eq. (12) can be nonzero, and stable. Then, the current density is given by

j=−ensv=−e2nsme(A−ℏ2e∇χ)(15)\mathbf{j}=-en_s\mathbf{v}=-\frac{e^2n_s}{m_e}\left(\mathbf{A}-\frac{\hbar}{2e}\nabla\chi\right) \tag*{(15)}

where nsn_s is the number density of electrons that flow with v\mathbf{v}; we explain the meaning of nsn_s, below.

In the wave function in Eq. (11), all NN electrons participate in the ‘χ\chi mode’ described by the factor e−i2χe^{-\frac{i}{2}\chi}. However, it is possible that the number can be different from NN. In other words, the wave function is given as a superposition of states with different number of electrons in the χ\chi mode,

Ψ(x1,…,xN,t)=C0Ψ00(x1,…,xN,t)+C1∑j1=1Ne−i2χ(rj1,t)Ψ01(x1,…,xN,t)+C2∑j1>j2e−i2[χ(rj1,t)+χ(rj2,t)]Ψ02(x1,…,xN,t)+C3∑j1>j2>j3e−i2[χ(rj1,t)+χ(rj2,t)+χ(rj3,t)]Ψ03(x1,…,xN,t)⋯⋯+CNe−i2[χ(r1,t)+⋯+χ(rN,t)]Ψ0N(x1,…,xN,t)(16)\begin{aligned} \Psi(x_1,\ldots,x_N,t)&=C_0\Psi_{00}(x_1,\ldots,x_N,t)\\ &\quad+C_1\sum_{j_1=1}^{N}e^{-\frac{i}{2}\chi(r_{j_1},t)}\Psi_{01}(x_1,\ldots,x_N,t)\\ &\quad+C_2\sum_{j_1>j_2}e^{-\frac{i}{2}[\chi(r_{j_1},t)+\chi(r_{j_2},t)]}\Psi_{02}(x_1,\ldots,x_N,t)\\ &\quad+C_3\sum_{j_1>j_2>j_3}e^{-\frac{i}{2}[\chi(r_{j_1},t)+\chi(r_{j_2},t)+\chi(r_{j_3},t)]}\Psi_{03}(x_1,\ldots,x_N,t)\\ &\quad\cdots\cdots\\ &+ C_N e^{-\frac{i}{2}[\chi(\boldsymbol{r}_1,t)+\cdots+\chi(\boldsymbol{r}_N,t)]}\Psi_{0N}(\boldsymbol{x}_1,\ldots,\boldsymbol{x}_N,t) \tag*{(16)} \end{aligned}

where C0,…,CNC_0,\ldots,C_N are constants; Ψ00,…,Ψ0N\Psi_{00},\ldots,\Psi_{0N} are currentless antisymmetric wave functions. Note that Ψ\Psi is antisymmetric with respect to the exchange of electrons. Ψ\Psi in Eq. (11) corresponds to the case C0=C1=⋯=CN−1C_0=C_1=\cdots=C_{N-1}, CN=1C_N=1 and Ψ0N=Ψ0\Psi_{0N}=\Psi_0.

Using this fluctuation of the number of electrons participating in the χ\chi mode, the system can reduce the total energy. A practical way to include this fluctuation is to use the modified Bogoliubov-de Gennes type equation, which is explained succinctly in A. The important ingredients in this formalism are the particle number conserving Bogoliubov operators γn\gamma_n and γn†\gamma_n^\dagger that describe the transitions between states with different numbers of electrons in the χ\chi mode. The superconducting electron density nsn_s is given as the number density of electrons participating in the χ\chi mode, ρχ\rho_\chi, in A.

From the viewpoint of the new theory, the BCS theory is the theory that takes into account the number fluctuation of electrons in the χ\chi mode as the number fluctuation of Cooper pairs; the convenient formalism to include such fluctuations is to use the particle number nonconserving Bogoliubov operators [31, 32]. In this respect, the BCS theory is a theory using the gauge symmetry approximation [33, 34]. The normal ground state solution corresponds to the case where C0=1C_0=1, C1=⋯=CN−1=CN=0C_1=\cdots=C_{N-1}=C_N=0 as explained by Bloch and Bohm [35], which yields ns=0n_s=0.

Let us examine the supercurrent generated by v\boldsymbol{v}. Since v\boldsymbol{v} is gauge invariant, we may choose A\boldsymbol{A} in the gauge ∇⋅A=0\nabla\cdot\boldsymbol{A}=0. Then, the conservation of the charge for the stationary case, ∇⋅j=0\nabla\cdot\boldsymbol{j}=0, yields

∇2χ=0(17)\nabla^2\chi=0 \tag*{(17)}

where the spatial dependence of nsn_s is neglected for simplicity. The solution to the above equation is characterized by an integer called the ‘winding number’,

w[χ]C=12π∮C∇χ⋅dr(18)w[\chi]_{C}=\frac{1}{2\pi}\oint_C\nabla\chi\cdot\mathrm{d}\boldsymbol{r} \tag*{(18)}

where CC is a loop in the space. If this number is non-zero, a persistent loop current flows. If a collection of such loop currents is stable, a macroscopic persistent current, ‘supercurrent’ will be realized. This supercurrent yields the flux quantization in the unit of h2e\frac{h}{2e} as observed in superconductors.

In the new theory, the appearance of χ\chi in Eq. (10) is attributed to the following two constrains on the solution of the Schrödinger equation:

  1. The conservation of local charge.

  2. The single-valuedness of the wave function as a function of the electron coordinates.

For the cuprate superconductivity case, the present author has been arguing that the spinor property of electrons plays an important role. It causes sign-change of the spin-function when spin is twisted one around for the excursion along the loop; then, in order to fulfill the second constraint, χ\chi appears on the coordinate function to compensate the spin function sign change. It can be shown that if we combine the above two constrains, we can obtain ∇χ\nabla\chi, which yields χ\chi with non-zero winding numbers [17, 28].

It is noteworthy that the collective plasma mode is obtained in the electron gas model by a constraint on the electron density, which is similar to the first constraint above [36]. In the new theory, the second one is added; it is needed to calculate nonzero current density using the wave function.

Anderson found the new collective mode in the BCS model due to the presence of Cooper pairs [37, 38]. In the new theory, this mode is replaced by the χ\chi mode. As the wave function in Eq. (16) indicates, it allows the single particle change of the number of electrons participating in it. This enables the Josephson tunneling by single-electron.

Plasma oscillation and screening

Extending the theoretical method that yields the plasma mode in the presence of Cooper pairs, Anderson found the new collective mode in the BCS model [37, 38]. This mode corresponds to the χ\chi mode in the new theory. In this section, we examine the effect of χ\chi mode in the plasma oscillation and screening. The effect we consider is that arising from the gradient of the chemical potential force. The non-zero winding number for χ\chi does not play any role here, although it is important in superconductivity.

We follow the Fetter and Walecka’s book [39] with adding the gradient of the chemical potential force: Let us consider an electron gas with electron density ρ\rho and the background compensating positive charge density ρ0\rho_0. If the electron density is slightly perturbed from the equilibrium value ρ0\rho_0

ρ(r,t)=ρ0+δρ(r,t)(19)\rho(\mathbf{r},t)=\rho_0+\delta\rho(\mathbf{r},t) \tag*{(19)}

an electric field E\mathbf{E} that satisfies

ϵ0∇⋅E=−eδρ(r,t)(20)\epsilon_0\nabla\cdot\mathbf{E}=-e\delta\rho(\mathbf{r},t) \tag*{(20)}

arise.

The forces acting on the electrons are the one from the electric field and another from the gradient of the chemical potential. Thus, the Newtonian equation yields

meddt(ρv)=−eρE−ρ∇μ(21)m_e\frac{d}{dt}(\rho\mathbf{v})=-e\rho\mathbf{E}-\rho\nabla\mu \tag*{(21)}

or

meρ0∂∂tv≈−eρ0E−ρ0∇μ(22)m_e\rho_0\frac{\partial}{\partial t}\mathbf{v}\approx-e\rho_0\mathbf{E}-\rho_0\nabla\mu \tag*{(22)}

Another relation is obtained from the equation of continuity,

∂ρ∂t+∇⋅(ρv)≈∂δρ∂t+ρ0∇⋅v=0(23)\frac{\partial\rho}{\partial t}+\nabla\cdot(\rho\mathbf{v})\approx\frac{\partial\delta\rho}{\partial t}+\rho_0\nabla\cdot\mathbf{v}=0 \tag*{(23)}

From Eqs. (20), (22), and (23), we obtain

∂2δρ∂t2=−ρ0∂∂t∇⋅v=eρ0me(∇⋅E+e−1∇2μ)=−Ωpl2δρ+ρ0me∇2μ(24)\begin{aligned} \frac{\partial^2 \delta\rho}{\partial t^2} &= -\rho_0 \frac{\partial}{\partial t}\nabla\cdot\mathbf{v} \\ &= \frac{e\rho_0}{m_e}\left(\nabla\cdot\mathbf{E} + e^{-1}\nabla^2\mu\right) \\ &= -\Omega_{pl}^2\delta\rho+ \frac{\rho_0}{m_e}\nabla^2\mu \tag*{(24)} \end{aligned}

where the plasma frequency is defined by

Ωpl=(e2ρ0ϵ0me)1/2(25)\Omega_{pl} = \left(\frac{e^2\rho_0}{\epsilon_0m_e}\right)^{1/2} \tag*{(25)}

If the gradient of the chemical potential force is neglected, the oscillation occurs with the frequency Ωpl\Omega_{pl}.

Let us consider the effect of the gradient of the chemical potential force. Using the free electron gas chemical potential μ0\mu_0, and assuming the deviation of μ\mu, δμ\delta\mu is very small, we have

μ=μ0+δμ,μ0=ℏ22me(3π2ρ0)2/3(26)\mu= \mu_0 + \delta\mu,\qquad\mu_0 = \frac{\hbar^2}{2m_e}(3\pi^2\rho_0)^{2/3} \tag*{(26)}

Then, the value of δμ\delta\mu to the order of δρ\delta\rho is calculated as

δμ=23μ0ρ0δρ(27)\delta\mu= \frac{2}{3}\frac{\mu_0}{\rho_0}\delta\rho \tag*{(27)}

Including δμ\delta\mu and using Eq. (27), Eq. (24) becomes

∂2δρ∂t2=−Ωpl2δρ+2μ03me∇2δρ(28)\frac{\partial^2\delta\rho}{\partial t^2} = -\Omega_{pl}^2\delta\rho+ \frac{2\mu_0}{3m_e}\nabla^2\delta\rho \tag*{(28)}

From this equation, the following dispersion relation is obtained for the Fourier component of δρ\delta\rho that oscillates with frequency Ωk\Omega_k and the wavenumber vector k\mathbf{k}

Ωk2=Ωpl2[1+(kkTF)2](29)\Omega_k^2 = \Omega_{pl}^2\left[1+\left(\frac{k}{k_{\mathrm{TF}}}\right)^2\right] \tag*{(29)}

where the Thomas-Fermi wavenumber is defied by

kTF=(3e2ρ02ϵ0μ0)1/2(30)k_{\mathrm{TF}} = \left(\frac{3e^2\rho_0}{2\epsilon_0\mu_0}\right)^{1/2} \tag*{(30)}

Although the constant factor for the second order term (it is 1 in the present calculation) in Eq. (29) is slightly different from the one obtained quantum mechanically using the random phase approximation (it is 910\frac{9}{10}; see [39]), essentially the same result is obtained.

Next, we consider the screening problem in the electron gas model. It is also a well-studied model; however, we consider it from the chemical potential view point. We follow the Pines’ book [40]. Let us place a static charge qq at the origin of the electron gas. The electric field generated by it is expressed using the scalar potential φ\varphi, E=−∇φ\mathbf{E}=-\nabla\varphi; it satisfies the following equation

−ϵ0∇2φ=qδ(r)−eδρind(31)-\epsilon_{0}\nabla^{2}\varphi=q\delta(\mathbf{r})-e\delta\rho_{\mathrm{ind}} \tag*{(31)}

where δρind\delta\rho_{\mathrm{ind}} denotes the induced charge density. We calculate δρind\delta\rho_{\mathrm{ind}} by employing the Thomas–Fermi model; then, the electron density is given by

ρ(r)=13π2[2meℏ2(ϵF+eφ)]3/2(32)\rho(\mathbf{r})=\frac{1}{3\pi^{2}}\left[\frac{2m_{e}}{\hbar^{2}}\left(\epsilon_{\mathrm{F}}+e\varphi\right)\right]^{3/2} \tag*{(32)}

where ϵF\epsilon_{\mathrm{F}} is the Fermi energy. From Eq. (13), the change of the chemical potential due to the addition of the point charge qq is given by

δμTF=e[φ+ℏ2e∂t(δχ)](33)\delta\mu_{\mathrm{TF}}=e\left[\varphi+\frac{\hbar}{2e}\partial_{t}(\delta\chi)\right] \tag*{(33)}

where the change of χ\chi is also included. Then, the electron density is given by

ρ(r)=13π2[2meℏ2(ϵF+δϵF+δμTF)]3/2,δϵF=ℏ2∂t(δχ)(34)\rho(\mathbf{r})=\frac{1}{3\pi^{2}}\left[\frac{2m_{e}}{\hbar^{2}}\left(\epsilon_{\mathrm{F}}+\delta\epsilon_{\mathrm{F}}+\delta\mu_{\mathrm{TF}}\right)\right]^{3/2},\qquad\delta\epsilon_{\mathrm{F}}=\frac{\hbar}{2}\partial_{t}(\delta\chi) \tag*{(34)}

Thus, δρind\delta\rho_{\mathrm{ind}} is approximately obtained as

δρind(r)=3ρ02μ0δμTF(35)\delta\rho_{\mathrm{ind}}(\mathbf{r})=\frac{3\rho_{0}}{2\mu_{0}}\delta\mu_{\mathrm{TF}} \tag*{(35)}

We rewrite the equation for φ\varphi in Eq. (31) to the equation for δμTF\delta\mu_{\mathrm{TF}}

∇2δμTF=−eqϵ0δ(r)+kTF2δμTF(36)\nabla^{2}\delta\mu_{\mathrm{TF}}=-\frac{eq}{\epsilon_{0}}\delta(\mathbf{r})+k_{\mathrm{TF}}^{2}\delta\mu_{\mathrm{TF}} \tag*{(36)}

It yields

δμTF=eq4πϵ0re−kTFr(37)\delta\mu_{\mathrm{TF}}=\frac{eq}{4\pi\epsilon_{0}r}e^{-k_{\mathrm{TF}}r} \tag*{(37)}

Thus, the scalar potential is obtained as

φ=q4πϵ0re−kTFr−e−1δϵF(38)\varphi=\frac{q}{4\pi\epsilon_{0}r}e^{-k_{\mathrm{TF}}r}-e^{-1}\delta\epsilon_{\mathrm{F}} \tag*{(38)}

Note that this contains the gauge ambiguity term −e−1δϵF-e^{-1}\delta\epsilon_{\mathrm{F}}. The value of ϵF+δϵF\epsilon_{\mathrm{F}}+\delta\epsilon_{\mathrm{F}} is obtained from the constraint that ρ\rho integrated over the system region is equal to the total number of electrons.

The force from the electric field E\mathbf{E} is given by

−eE=e∇φ=∇eq4πϵ0re−kTFr(39)-e\mathbf{E}=e\nabla\varphi=\nabla\frac{eq}{4\pi\epsilon_{0}r}e^{-k_{\mathrm{TF}}r} \tag*{(39)}

and that from the gradient of the chemical potential is

−∇δμTF=−∇eq4πϵ0re−kTFr(40)-\nabla\delta\mu_{\mathrm{TF}}=-\nabla\frac{eq}{4\pi\epsilon_{0}r}e^{-k_{\mathrm{TF}}r} \tag*{(40)}

The sum of them is zero, indicating that the force acting on the electrons is zero. Actually, it is a consequence of the fact that E\mathbf{E} is described only by the scalar potential. In this case, Eq. (13) yields the compensation of the electrostatic force and gradient of the chemical potential force if ∂tχ\partial_{t}\chi dose not have the spatial dependence.

It is important to note that the force from the electrostatic field and that from the gradient of the chemical potential tend to compensate. This will occur in the situation where a steady electric current flows.

Differentiation of normal and superconducting metals

Before the advent of high transition temperature cuprate superconductors [41], it was believed that the BCS theory had solved all major problems of superconductivity [21]. In this theory, the superconducting state is explained as the state stabilized by electron-pair formation; the temperature where the electron pair is formed corresponds to the superconducting transition temperature. Thus, the energy gap formed by the electron-pair formation was considered as the theoretical hallmark of superconductors, and methods to accurately obtain the energy gap formation temperature have been developed.

However, the elucidation of the mechanism of superconductivity in the cuprates has not been successful based on the BCS theory. A theory indicates that the superconducting transition temperature corresponds to the stabilization temperature of nano-sized loop currents [42] and a simulation based on the spin-vortex-induced loop current model has yielded the result that supports this claim [43]. Besides, revisiting of experimental facts of superconductivity has found several problems even in the standard theory based on the BCS one:

  1. The standard theory relies on the use of particle number non-conserving formalism to explain the U(1)U(1) gauge symmetry breaking although superconductivity occurs in an isolated system where the particle-number is conserved [33].

  1. The superconducting carrier mass obtained by the London moment experiment is the free electron mass mem_e, although the standard theory prediction is the effective mass m∗m^{*} of the normal state [44]. (The London moment is a phenomenon associated with the rotating superconductor. A magnetic field is generated inside of it , and

this magnetic field is called the ‘London magnetic field’ and the magnetic moment associated with it is the ‘London moment’, which has been measured.)

  1. The correct centripetal force in the London moment experiment cannot be obtained by the standard theory [44, 45].

  1. The reversible superconducting-normal phase transition in a magnetic field cannot be explained by the standard theory [46].

  1. The dissipative quantum phase transition in a Josephson junction system predicted by the standard theory is absent [25].

  1. The so-called ‘quasiparticle poisoning problem’ indicates the existence a large amount of excited single electrons in Josephson junction systems for qubits, obtaining the observed ratio of their number to the Cooper pair number 10−9∼10−510^{-9} \sim10^{-5} in disagreement with the standard theory ratio 10−5210^{-52} [26, 27].

The existence of the above problems indicates the need for serious revisions of the superconductivity theory. A theory that encompasses the BCS theory has been developed by the present author using the Berry connection from many-electron wave functions [16, 17] (A brief explanation of the relation between the new theory and the BCS one is given in A). It resolves the above problems. In the following, we explain how normal and superconducting metals are differentiated by the new theory.

First, we consider a general case where the Berry connection from many-body wave functions is composed of contributions from a set of states {Ψj}\{\Psi_j\} whose occupation probability is given by the Boltzmann distribution,

AMB=∑jpjAΨjMB,pj=e−EjkBT∑je−EjkBT(41)\mathbf{A}^{\mathrm{MB}}=\sum_j p_j\mathbf{A}_{\Psi_j}^{\mathrm{MB}},\qquad p_j=\frac{e^{-\frac{E_j}{k_{\mathrm{B}}T}}}{\sum_j e^{-\frac{E_j}{k_{\mathrm{B}}T}}} \tag*{(41)}

where AΨjMB\mathbf{A}_{\Psi_j}^{\mathrm{MB}} is the Berry connection from many-body wave functions for the wave function Ψj\Psi_j [11, 47], and kBk_{\mathrm{B}} is Planck’s constant. We rewrite AMB\mathbf{A}^{\mathrm{MB}} as

AMB=−12∇χ(42)\mathbf{A}^{\mathrm{MB}}=-\frac{1}{2}\nabla\chi \tag*{(42)}

Here, we do not assume χ\chi to be an angular variable with period 2π2\pi as in superconductors since it may be an average of AΨjMB\mathbf{A}_{\Psi_j}^{\mathrm{MB}} from many Ψj\Psi_j’s. It becomes an angular variable with period 2π2\pi when only one Ψj\Psi_j is involved. In any case, we express v\mathbf{v} in the form given in Eq. (12), and μ\mu in Eq. (13).

The supercurrent in Eq. (15) explains a number of characteristics observed in superconductors as follows:

  1. Persistent current: It can be explained as due to the presence of χ\chi with stable non-zero w[χ]Cw[\chi]_{\mathrm{C}} [47].

  1. Meissner effect: If ∇×\nabla\times is operated on the both sides of Eq. (15), ∇2B=e2nmeB\nabla^2\mathbf{B}=\frac{e^2n}{m_e}\mathbf{B} is obtained. This is the equation obtained by London for explaining the Meissner effect.

  1. Flux quantization:

If we take a loop CC in Eq. (18) along a path inside a ring-shaped superconductor where B=j=0\mathbf{B}=\mathbf{j}=0, we obtain ∮CA⋅dr=ℏ2ew[χ]C\oint_C \mathbf{A}\cdot d\mathbf{r}=\frac{\hbar}{2e}w[\chi]_C. This explains the observed flux quantization.

London moment

Consider a rotating superconductor with an angular velocity ω\boldsymbol{\omega}. The velocity field inside the superconductor is v=ω×r\mathbf{v}=\boldsymbol{\omega}\times\mathbf{r}. Substitute this v\mathbf{v} in Eq. (12), and operate ∇×\nabla\times on the both sides of it, the relation 2ω=2emeB2\boldsymbol{\omega}=\frac{2e}{m_e}\mathbf{B} is obtained. This indicates that the magnetic field B=2meeω\mathbf{B}=\frac{2m_e}{e}\boldsymbol{\omega} is created inside the superconductor. This is the field that has been observed, experimentally.

The normal conductor is the one without stable non-trivial χ\chi that satisfies Eq. (17). In this case, the gauge ambiguity of A\mathbf{A} in Eq. (15) is not absorbed by χ\chi. If j\mathbf{j} is calculated as an average value over different A\mathbf{A}’s allowed by the gauge ambiguity, it will result in zero.

Energy flow for a metallic wire connected to a battery

Now we examine an old dissipation phenomenon, the Joule heating by electric current in a metallic wire connected to an electrical battery, with including the Berry connection from many-body wave functions. The dissipation is believed to arise from the thermal fluctuation of the electric field. We use the following Langevin equation [48, 49] to take into account the fluctuation

medvdt=−meτv+η(43)m_e\frac{d\mathbf{v}}{dt}=-\frac{m_e}{\tau}\mathbf{v}+\boldsymbol{\eta} \tag*{(43)}

where τ\tau is the relaxation time; the electron mass is simply taken to be the free electron mass, but the effective mass m∗m^* will be more appropriate. The thermal average of the noise η=(η1,η2,η3)\boldsymbol{\eta}=(\eta_1,\eta_2,\eta_3) satisfies the following relations,

⟨ηi(t)⟩=0,⟨ηi(t)ηj(t′)⟩=2mekBTτδijδ(t−t′)(44)\langle\eta_i(t)\rangle=0,\qquad\langle\eta_i(t)\eta_j(t')\rangle=\frac{2m_e k_B T}{\tau}\delta_{ij}\delta(t-t') \tag*{(44)}

where kBTk_B T arises from the assumption that the thermal average of velocity v=(v1,v2,v3)\mathbf{v}=(v_1,v_2,v_3) satisfies

12me⟨vi2(t)⟩=12kBT(45)\frac{1}{2}m_e\langle v_i^2(t)\rangle=\frac{1}{2}k_B T \tag*{(45)}

according to the Maxwell velocity distribution. The noise arising from the fluctuating electric field is the cause of the dissipation.

Now we consider the case where a force is added. The Langevin equation in Eq. (43) with an additional force F\mathbf{F} on the right-hand side has the following solution

⟨v(t)⟩=τmeF(1−e−tτ)+⟨v(0)⟩e−tτ(46)\langle\mathbf{v}(t)\rangle=\frac{\tau}{m_e}\mathbf{F}\left(1-e^{-\frac{t}{\tau}}\right)+\langle\mathbf{v}(0)\rangle e^{-\frac{t}{\tau}} \tag*{(46)}

which yields the final velocity

⟨v(∞)⟩=τmeF(47)\langle\mathbf{v}(\infty) \rangle= \frac{\tau}{m_{\mathrm{e}}}\mathbf{F} \tag*{(47)}

The situation relevant to the Ohm’s law is the one with the above stationary current, and we identify it to the drift velocity

vd=τmeF(48)\mathbf{v}_{\mathrm{d}} = \frac{\tau}{m_{\mathrm{e}}}\mathbf{F} \tag*{(48)}

When the stationary drift velocity is established, the velocity fluctuation occurs around vd\mathbf{v}_{\mathrm{d}}. This situation may be described by modifying Eq. (43) as

med(v−vd)dt=−meτ(v−vd)+η′,η′=η+F(49)m_{\mathrm{e}}\frac{\mathrm{d}(\mathbf{v}-\mathbf{v}_{\mathrm{d}})}{\mathrm{d}t} = -\frac{m_{\mathrm{e}}}{\tau}(\mathbf{v}-\mathbf{v}_{\mathrm{d}})+\boldsymbol{\eta}^{\prime}, \qquad\boldsymbol{\eta}^{\prime}=\boldsymbol{\eta}+\mathbf{F} \tag*{(49)}

with

⟨ηi′(t)⟩F=0,⟨ηi′(t)ηj′(t′)⟩F=2mekBTτδijδ(t−t′)(50)\langle\eta_i^{\prime}(t)\rangle_{\mathbf{F}}=0,\qquad \langle\eta_i^{\prime}(t)\eta_j^{\prime}(t^{\prime})\rangle_{\mathbf{F}} =\frac{2m_{\mathrm{e}}k_{\mathrm{B}}T}{\tau}\delta_{ij}\delta(t-t^{\prime}) \tag*{(50)}

where ⟨⋯ ⟩F\langle\cdots\rangle_{\mathbf{F}} indicates the thermal average in the presence of F\mathbf{F}. Here, we assume that the effect of the presence of F\mathbf{F} on τ\tau is negligible, and the same τ\tau in Eq. (43) can be used in Eq. (49). The center of the noise is shifted, and the noise from it is denoted by η′\boldsymbol{\eta}^{\prime}.

From Eqs. (49) and (50), the following relation is obtained

⟨η⟩F=−F(51)\langle\boldsymbol{\eta}\rangle_{\mathbf{F}}=-\mathbf{F} \tag*{(51)}

It is the relation between the shift of the fluctuation force and applied one.

Since the noise arises from the fluctuating electric field, the shift is the induced electric field given by

Eind=−1e⟨η⟩F(52)\mathbf{E}_{\mathrm{ind}}=-\frac{1}{e}\langle\boldsymbol{\eta}\rangle_{\mathbf{F}} \tag*{(52)}

The steady current flows under the influence of the forces F\mathbf{F}, Eind\mathbf{E}_{\mathrm{ind}}, and the fluctuating one η′\boldsymbol{\eta}^{\prime}.

Now we consider the Joule heating by the Drude model. In the textbook explanation based on the Drude theory, the electric current is obtained using Eq. (48) with F=−eEDrude\mathbf{F}=-e\mathbf{E}_{\mathrm{Drude}}. The induced field in Eq. (52) is not taken into account, thus, the electric field is EDrude\mathbf{E}_{\mathrm{Drude}}. The voltage VV in Eq. (1) is given by

V=−∫r1r2EDrude⋅dr(53)V=-\int_{\mathbf{r}_1}^{\mathbf{r}_2}\mathbf{E}_{\mathrm{Drude}}\cdot\mathrm{d}\mathbf{r} \tag*{(53)}

where r1\mathbf{r}_1 and r2\mathbf{r}_2 are points where the battery is connected.

The Poynting vector EDrude×H\mathbf{E}_{\mathrm{Drude}}\times\mathbf{H} points radially inwards to the wire (see Figure 1), supplying energy for the Joule heat EDrude⋅j=σ−1j2\mathbf{E}_{\mathrm{Drude}}\cdot\mathbf{j}=\sigma^{-1}j^2 [4]. This is strange since it is sensible to consider the energy is supplied by the battery by the current through the wire.

Figure 1. A part of a metallic wire connected to a battery. The effect of the battery connection is the appearance of an electric field EDrude\mathbf{E}_{\mathrm{Drude}} in the Drude model; this gives rise to the force that accelerates conduction electrons. The electric current II flows in the same direction as EDrude\mathbf{E}_{\mathrm{Drude}} in the wire, generating a circumferential magnetic field B\mathbf{B} around the wire. The Poynting vector EDrude×B/μ0\mathbf{E}_{\mathrm{Drude}} \times\mathbf{B}/\mu_0 points radially inwards to the wire. Thus, radiation enters into the wire. The radiation energy entered is consumed as the Joule heat j⋅EDrude\mathbf{j} \cdot\mathbf{E}_{\mathrm{Drude}} in the wire.

Let us now assume that the voltage VV actually arises from the chemical potential difference realized by some chemical processes inside the battery. In other words, the electromotive force is generated by chemical reactions on the electrodes, where electrons are pushed-in from one of the electrodes, and pulled-out from the other. If the wire is made of a simple metal, the Fermi level increases on the number density increased connection point, and decreases on the other (see Eq. (26)). This difference gives rise to the electromotive force that generates the current. Then, the voltage VV is given by

V=−e−1∫r1r2∇μb⋅dr=−μb(r2)−μb(r1)e(54)V=-e^{-1}\int_{r_1}^{r_2}\nabla\mu_b\cdot d\mathbf{r}=-\frac{\mu_b(r_2)-\mu_b(r_1)}{e} \tag*{(54)}

where μb\mu_b is the chemical potential whose origin is the battery.

We consider that the situation given by the Langevin equation in Eq. (49) is relevant, rather than the one described by the Langevin equation in Eq. (43). In this case, as in the screening problem considered in § 3, the balance of the electric force and the gradient of the chemical potential force is taken into account. By taking F=−∇μb\mathbf{F}=-\nabla\mu_b, and using Eqs. (48), (51), and (52), the following relations are obtained,

j=e−1σ∇μb,Eind=−e−1∇μb(55)\mathbf{j}=e^{-1}\sigma\nabla\mu_b,\qquad\mathbf{E}_{\mathrm{ind}}=-e^{-1}\nabla\mu_b \tag*{(55)}

where σ\sigma is given by

σ=e2neτme(56)\sigma=\frac{e^2n_e\tau}{m_e} \tag*{(56)}

with nen_e being the conduction electron density. The direction of Eind\mathbf{E}_{\mathrm{ind}} is opposite to the current direction (see Figure 2); thus, the direction of the Poynting vector Eind×B/μ0\mathbf{E}_{\mathrm{ind}}\times\mathbf{B}/\mu_0 points radially outwards from the wire.

Figure 2. The same as in Figure 1 but the battery connection gives rise to the chemical potential gradient −∇μb-\nabla\mu_b inside the wire. An electric field Eind\mathbf{E}_{\mathrm{ind}} is generated to counteract the force from the gradient of −∇μb-\nabla\mu_b; due to the balance of the two forces, the electrons perform stationary translational motion with the velocity vd\mathbf{v}_d. The Poynting vector Eind×B/μ0\mathbf{E}_{\mathrm{ind}}\times\mathbf{B}/\mu_0 points radially outwards from the wire, thus, radiation is emitted. The radiation energy emitted is equal to the Joule heat j⋅Eind\mathbf{j}\cdot\mathbf{E}_{\mathrm{ind}}, which is the work done by the battery.

If we follow the typical textbook calculation of the Poynting theorem energy flow for the above situation, it goes as follows: Let us assume the part of the wire depicted is a cylindrical shape with circular cross sectional area of radius rr and length ℓ\ell. Then, a magnetic field with the magnitude H=I2πrH=\frac{I}{2\pi r} exists at the surface of the wire. The magnitude of the Poynting vector Eind×H\mathbf{E}_{\mathrm{ind}}\times\mathbf{H} is IEind2πr\frac{IE_{\mathrm{ind}}}{2\pi r}. By integrating it over the side surface of the cylinder, the total energy flow is obtained as

∫SdS⋅(Eind×H)=IEindℓ(57)\int_S d\mathbf{S}\cdot(\mathbf{E}_{\mathrm{ind}}\times\mathbf{H})=IE_{\mathrm{ind}}\ell \tag*{(57)}

where SS is the surface of the cylinder; the Poynting vector goes out only through the side surface of the cylinder with area 2πr2\pi r. Since Eind\mathbf{E}_{\mathrm{ind}} points in the opposite direction to j\mathbf{j}, we have

∫Vd3r j⋅Eind=−∫Vd3r jEind=−IEindℓ(58)\int_{\mathcal{V}}d^3r\,\mathbf{j}\cdot\mathbf{E}_{\mathrm{ind}}=-\int_{\mathcal{V}}d^3r\,jE_{\mathrm{ind}}=-IE_{\mathrm{ind}}\ell \tag*{(58)}

where V\mathcal{V} is the volume of the cylinder, and relation I=jπr2I=j\pi r^2 is used. Then, the relation in Eq. (3) is satisfied since the electromagnetic field energy in the wire is constant in the stationary situation. The term with j⋅Eind\mathbf{j}\cdot\mathbf{E}_{\mathrm{ind}} has the minus sign, indicating that it is the work done by the battery. Thus, the Poynting theorem now states that the energy for the emitted radiation is supplied by the battery through the wire. This is a sensible result. This emitted radiation is the Joule heat; when it is absorbed by some materials it will cause heating.

Actually, the Drude model description for the energy flow is relevant for the optical conductivity case. In this case the electric field is applied from outside of the sample as radiation, and the relation in Eq. (4) is valid. The absorption of the radiation occurs in the sample, which can be calculated usually using the Fermi’s golden rule formalism of quantum mechanics; this will correspond to the radiation energy consumption.

The conductivity by the battery connection is, in a sense, the reverse process of the optical conductivity case. For the battery connection case, collisions among the conduction electrons and those between the ion cores and the electrons generate quantum mechanical excited states that correspond to the exited states created by the absorption of the radiation in the optical conductivity case. The emission of radiation is due to the relaxation of the excited states, which is the reverse process of the excitation by the applied radiation for the optical conductivity case. The generation of the induced field Eind\mathbf{E}_{\mathrm{ind}} and resulting emission of radiations by the Poynting vector may be considered as classical approximations for the above quantum mechanical processes.

The differences between metals and semiconductors cannot be described by the present method. It requires microscopic quantum mechanical calculations, which can differentiate electronic excitations in metals and semiconductors.

Energy flow during discharging of a capacitor

Let us consider the discharging of a capacitor. In the beginning, the charges +Q0+Q_0 and −Q0-Q_0 are on each electrodes of the capacitor, respectively (see Figure 3), and the chemical potential difference Q0/CQ_0/C exists across the electrodes, where CC is the capacitance. of the capacitor. This chemical potential difference is due to the difference of the electron number density on each electrode. We may consider that this situation is realized first by connecting the capacitor to a battery of the voltage Q0/CQ_0/C, and then, by disconnecting the battery after the charging that establishes the voltage balance between the battery and the capacitor.

Image not extracted

Figure 3. A capacitor of capacitance CC with charge QQ stored and a wire with resistance RR that shunts the capacitor. During discharging QQ decreases as Q(t)=Q0e−1RCtQ(t)=Q_0e^{-\frac{1}{RC}t}, where Q0Q_0 is the initial charge at t=0t=0. The electric current I=dQdtI=\frac{dQ}{dt} flows after t=0t=0. We assume the balance of the voltage given in Eq. (59) is maintained in the quasi-stationary current situation. The electric field energy originally stored between the electrodes is Q022C\frac{Q_0^2}{2C}, which is emitted as the radiation from the wire.

The energy stored as the energy of the electric field E\mathbf{E} between the electrodes. The balance of the voltage between the chemical potential difference between the two electrodes and the electromagnetic field E\mathbf{E} in the space between the electrodes is given by

V=−∫12E⋅dℓ=1−e∫12∇μc⋅dℓ(59)V=-\int_{1}^{2}\mathbf{E}\cdot d\ell=\frac{1}{-e}\int_{1}^{2}\nabla\mu_c\cdot d\ell \tag*{(59)}

where ‘1’ and ‘2’ indicate points in the electrodes; the subscript cc in the chemical potential μc\mu_c indicates that it is the capacitor origin.

Now, we shunt the capacitor by the connection of a resistive wire. We assume a quasi-stationary current flow. It flows due to the chemical potential difference between the electrodes. The capacitor acts as the battery that produces the gradient of the chemical potential ∇μc\nabla\mu_c. The decrease of the electric field E\mathbf{E} occurs due to the decrease of the charge stored on the capacitor, thus, the reduction of the electric field energy stored also occurs. A typical textbook explanation goes as follows: The time variation of E\mathbf{E} generates a magnetic field B\mathbf{B} by the displacement current, and the Poynting vector E×H\mathbf{E}\times\mathbf{H} directing radially outwards from the capacitor is generated. The reduction of the electric field energy occurs due to the outward radiation from the space between the electrodes.

On the other hand, the present explanation is as follows: A\mathbf{A} and φ\varphi are fundamental physical quantities instead of E\mathbf{E} and B\mathbf{B}; then, E\mathbf{E} and B\mathbf{B} are two disguises of A\mathbf{A} and φ\varphi. If we adopt the gauge ∇⋅A=0\nabla\cdot\mathbf{A}=0, the vector potential is given as follows [50]

A(r,t)=μ04π∫d3r′j(r′,t−∣r−r′∣/c)∣r−r′∣(60)\mathbf{A}(\mathbf{r},t)=\frac{\mu_0}{4\pi}\int d^3r'\frac{\mathbf{j}\left(\mathbf{r}',t-\lvert\mathbf{r}-\mathbf{r}'\rvert/c\right)}{\lvert\mathbf{r}-\mathbf{r}'\rvert} \tag*{(60)}

This indicates that only the true current j\mathbf{j} generates A\mathbf{A}; thus, the generation of B\mathbf{B} by the displacement current does not occur. It may be worth pointing out that Maxwell obtained a similar formula (the vector potential is called the ‘electromagnetic momentum’) [51]; he included the displacement current in j\mathbf{j}, which is incorrect.

The energy flow occurs through the wire by the current generated by the chemical potential gradient generated by the capacitor that acts as a battery. Applying Eq. (58) for the wire in Figure 3 with neglecting the effect of bending of the wire, the rate of the work done by the capacitor is calculated as

∫Vd3r j⋅E=I(t)V(t)(61)\int_{\mathcal{V}}d^3r\,\mathbf{j}\cdot\mathbf{E}=I(t)V(t) \tag*{(61)}

where V(t)V(t) is the voltage on the capacitor, I(t)I(t) is the current flowing in the wire, and now V\mathcal{V} is the volume of the shunt wire. The total work done by the capacitor is obtained by integrating Eq. (61) in the time interval 0≤t≤∞0 \le t \le\infty as

∫0∞I(t)V(t) dt=∫0∞dQ(t)dtQ(t)C dt=−Q022C(62)\int_{0}^{\infty} I(t)V(t)\,dt = \int_{0}^{\infty} \frac{dQ(t)}{dt}\frac{Q(t)}{C}\,dt = -\frac{Q_0^2}{2C} \tag*{(62)}

where Q(t)=Q0e−t/RCQ(t)=Q_0e^{-t/RC} is used. The total work capacitor done is equal to the energy stored originally as the electric field. As shown in Eq. (57), this energy is equal to the energy irradiated from the wire as the Joule heat.

Josephson junction

Now we consider the supercurrent flow through the Josephson junction. The Josephson junction is schematically depicted in Figure 4. The current through it is given by

Image not extracted

Figure 4. Josephson junction composed of two superconducting electrodes (S1S_1 and S2S_2) and an insulator (gray region) between them. This structure inherently has a capacitor contribution due to the presence of two electrodes facing each other with the insulator between them.

I=Icsin⁡ϕ(63)I=I_c\sin\phi \tag*{(63)}

where IcI_c is the critical current of the junction and ϕ\phi is the superconducting phase difference between the two superconductors [23]. When a finite voltage exits across the Josephson junction, the time variation of ϕ\phi given by

dϕdt=2eVℏ(64)\frac{d\phi}{dt}=\frac{2eV}{\hbar} \tag*{(64)}

occurs, where VV is the voltage across the junction. This formula has been confirmed, experimentally [52]. It was first obtained by Josephson [22], and also derived by Feynman in his textbook [23]. By assuming the Cooper-pair tunneling, ϕ\phi is given by

ϕ=2eℏ∫12(A−ℏ2e∇χ)⋅dr(65)\phi=\frac{2e}{\hbar}\int_{1}^{2}\left(\mathbf{A}-\frac{\hbar}{2e}\nabla\chi\right)\cdot d\mathbf{r} \tag*{(65)}

where ‘1’ and ‘2’ are points in the two superconductors S1S_1 and S2S_2 of the junction, respectively. Note that (A−ℏ2e∇χ)\left(\mathbf{A}-\frac{\hbar}{2e}\nabla\chi\right) is gauge invariant, and 2e2e in the factor 2eℏ\frac{2e}{\hbar} is the Cooper-pair charge.

A Josephson junction is composed of an electron transfer contribution and a capacitor contribution as depicted with circuit elements in Figure 5. The capacitor contribution arises from the structure of two facing electrodes as seen in Figure 4. The tunneling electron goes through the transfer part with experiencing the electric field between the superconducting electrodes giving rise to the capacitor contribution. This capacitor contribution is lacking in the Josephson’s and textbook derivations [22, 23].

Image not extracted

Figure 5. A Josephson junction composed of a capacitor with capacitance CJC_J and an electron transfer part with critical current IcI_c.

Let us calculate the time derivative of Eq. (65) with including the capacitor contribution. The result is

dϕdt=2eℏ∫12(∂tA−ℏ2e∇χ)⋅dr=2eℏ∫12(−E−e−1∇μ)⋅dr(66)\frac{d\phi}{dt}=\frac{2e}{\hbar}\int_{1}^{2}\left(\partial_t\mathbf{A}-\frac{\hbar}{2e}\nabla\chi\right)\cdot d\mathbf{r} =\frac{2e}{\hbar}\int_{1}^{2}\left(-\mathbf{E}-e^{-1}\nabla\mu\right)\cdot d\mathbf{r} \tag*{(66)}

where μ\mu in Eq. (13) and E=−∇φ−∂tA\mathbf{E}=-\nabla\varphi-\partial_t\mathbf{A} are used. The voltage balance between the gradient of the chemical potential part and the electric field part gives rise to the relation in Eq. (59). Then, we have dϕdt=4eVℏ\frac{d\phi}{dt}=\frac{4eV}{\hbar}, which disagrees with Eq. (64). This discrepancy is resolved if we consider that the electron transfer occurs singly, not pair-wise. This indicates that the factor 2eℏ\frac{2e}{\hbar} in Eq. (65) is actually eℏ\frac{e}{\hbar}. This situation was considered by the present author, previously [53]. It is also worth noting that Eq. (16) indicates that the single-electron change in the superconducting condensate in the new theory.

Since the pairing stabilization exists in superconductors, there are situations where subsequent single electron transfers observed as the pair-wise transfers. Thus, this result does not contradict experiment. It is noteworthy that it explains the so-called ‘quasiparticle poisoning problem’: In quantum electronic circuits with Josephson junctions, a large amount of excited single electrons in Josephson junctions is observed, with the observed ratio of their number to the Cooper pair number 10−9∼10−510^{-9} \sim10^{-5} in disagreement with the standard theory ratio 10−5210^{-52}[26, 27]; this may be the consequence of abundant single-electron transfer.

Planckian dissipation

We consider another superconductivity related problem. It is so-called, the ‘Planckian dissipation’ problem. The dissipations with the relaxation time of the Planckian time order have been observed in anomalous metals [18, 19, 20], including the high transition temperature cuprate superconductors. It is believed that the elucidation of it is the key to understand the cuprate superconductivity. We examine this problem with including the fluctuation of the gradient of the chemical potential force.

First, we consider the fluctuation of −e∇φ-e\nabla\varphi term arising from −∇μ-\nabla\mu, where μ\mu is given in Eq. (13). We assume that the fluctuation from −e∇φ-e\nabla\varphi can be mimicked by the force from an electric field eEe\mathbf{E}. The motion of electrons is assumed in the xx direction. Then, we have

τ=2mekBTe2⟨Ex(t)Ex(t′)⟩δ(t−t′)≈2mekBTe2⟨Ex(t)Ex(t)⟩1τE(67)\tau= \frac{2m_{\mathrm e}k_{\mathrm B}T}{e^2\langle E_x(t)E_x(t')\rangle}\delta(t-t') \approx\frac{2m_{\mathrm e}k_{\mathrm B}T}{e^2\langle E_x(t)E_x(t)\rangle}\frac{1}{\tau_{\mathrm E}} \tag*{(67)}

where ExE_x is the xx component of the fluctuating electric field. The effect of the delta function is replaced by a temperature independent constant τE\tau_{\mathrm E} with the dimension of time. This is the crudest approximation based on the following: ⟨Ex(t)Ex(t′)⟩\langle E_x(t)E_x(t')\rangle is assumed to be non-zero in the time interval τE\tau_{\mathrm E} with a typical value ⟨Ex(t)Ex(t)⟩\langle E_x(t)E_x(t)\rangle, and the delta function is replaced by a rectangle with the value τE−1\tau_{\mathrm E}^{-1} in the time-interval τE\tau_{\mathrm E}.

We may express the electric field Ex(t)E_x(t) using the creation and annihilation operators for the photon with energy ℏωn\hbar\omega_n, a^kn†\hat a^\dagger_{k_n} and a^kn\hat a_{k_n}, respectively

Ex(t)=i∑nℏωn2ϵ0LxLyLz(a^kne−i(ωnt−knx)−a^kn†ei(ωnt−knx))(68)E_x(t) = i\sum_n\sqrt{\frac{\hbar\omega_n}{2\epsilon_0L_xL_yL_z}}\left(\hat a_{k_n}e^{-i(\omega_nt-k_nx)}-\hat a^\dagger_{k_n}e^{i(\omega_nt-k_nx)}\right) \tag*{(68)}

LxLyLzL_xL_yL_z is the volume of a rectangular cuboid with length LxL_x in the xx direction, width LyL_y in the yy direction, and height LzL_z in the zz direction; the wave number knk_n is given by kn=2πLxnk_n=\frac{2\pi}{L_x}n, (nn is an integer), and ωn\omega_n by ωn=c∣kn∣\omega_n=c|k_n|. Then, ⟨Ex(t)Ex(t)⟩\langle E_x(t)E_x(t)\rangle is estimated as

⟨Ex(t)Ex(t)⟩=∑nℏωn2ϵ0LxLyLz⟨2a^kn†a^kn+1⟩≈1πϵ0LyLz∫0∞dk ℏcke−ℏckkBT=ℏcπϵ0LyLz(kBTℏc)2(69)\begin{aligned} \langle E_x(t)E_x(t)\rangle&= \sum_n\frac{\hbar\omega_n}{2\epsilon_0L_xL_yL_z}\langle2\hat a^\dagger_{k_n}\hat a_{k_n}+1\rangle\\ &\approx\frac{1}{\pi\epsilon_0L_yL_z}\int_0^\infty dk\,\hbar ck e^{-\frac{\hbar ck}{k_{\mathrm B}T}} \\ &= \frac{\hbar c}{\pi\epsilon_0L_yL_z}\left(\frac{k_{\mathrm B}T}{\hbar c}\right)^2 \tag*{(69)} \end{aligned}

where the sum ∑n\sum_n is replaced by Lx2π∫dk\frac{L_x}{2\pi}\int dk and the approximation ⟨2a^kn†a^kn+1⟩≈2e−ℏc∣kn∣kBT\langle2\hat{a}^{\dagger}_{k_n}\hat{a}_{k_n}+1\rangle\approx2e^{-\frac{\hbar c|k_n|}{k_BT}} is used, where e−ℏc∣kn∣kBTe^{-\frac{\hbar c|k_n|}{k_BT}} is the Boltzmann factor. Then, τ\tau is given by

τ≈meLyLz2αℏτEℏkBT(70)\tau\approx\frac{m_eL_yL_z}{2\alpha\hbar\tau_E}\frac{\hbar}{k_BT} \tag*{(70)}

where α≈1137\alpha\approx\frac{1}{137} is the fine structure constant. There are unknown parameters, LyL_y, LzL_z and τχ\tau_\chi, but τ\tau exhibits the dependence on ℏkBT\frac{\hbar}{k_BT}, a signature of the Planckian dissipation. Due to the presence of α−1\alpha^{-1}, however, the above estimate is too large to explain the observed τ∼ℏkBT\tau\sim\frac{\hbar}{k_BT} relation if unknown parameters are in the order of 1 in the atomic units.

Let us now estimate τ\tau using the fluctuation of −ℏ2∂t(∇χ)-\frac{\hbar}{2}\partial_t(\nabla\chi) from −∇μ-\nabla\mu, the ‘gauge fluctuation’; we assume it is associated with the fluctuation of the winding number in Eq. (18). First, we write τ\tau as

τ=2mekBTℏ24⟨∂t∂xχ(t)∂t∂xχ(t′)⟩δ(t−t′)≈2mekBTℏ24⟨∂t∂xχ(t)∂t∂xχ(t)⟩1τχ(71)\tau=\frac{2m_ek_BT}{\frac{\hbar^2}{4}\langle\partial_t\partial_x\chi(t)\partial_t\partial_x\chi(t')\rangle}\delta(t-t')\approx\frac{2m_ek_BT}{\frac{\hbar^2}{4}\langle\partial_t\partial_x\chi(t)\partial_t\partial_x\chi(t)\rangle}\frac{1}{\tau_\chi} \tag*{(71)}

where the effect of the delta function is replaced by a temperature independent constant τE\tau_E with the dimension of time as in the previous case.

We also perform the following approximation

∂t∂xχ(t)≈∂xχ(t+τχ)−∂xχ(t)τχ≈∂xχ(t+τχ)τχ, or −∂xχ(t)τχ(72)\partial_t\partial_x\chi(t)\approx\frac{\partial_x\chi(t+\tau_\chi)-\partial_x\chi(t)}{\tau_\chi}\approx\frac{\partial_x\chi(t+\tau_\chi)}{\tau_\chi},\ \text{or}\ -\frac{\partial_x\chi(t)}{\tau_\chi} \tag*{(72)}

This approximation is based on the assumption that the abrupt change of ∂xχ\partial_x\chi occurs in an average interval of τχ\tau_\chi, and the abrupt change is due to the creation or annihilation of the vortices of ∇χ\nabla\chi that is approximated as ∂xχ=0\partial_x\chi=0 in the before or in the after. We assume such abrupt change occurs as the consequence of the fluctuation of the winding number in Eq. (18).

Using the above approximation and assuming that the vortices exit in the xyxy plane, we have

⟨∂t∂xχ(t)∂t∂xχ(t)⟩≈1τχ2⟨∂xχ(t)∂xχ(t)⟩≈12τχ2⟨(∇χ(t))2⟩(73)\langle\partial_t\partial_x\chi(t)\partial_t\partial_x\chi(t)\rangle\approx\frac{1}{\tau_\chi^2}\langle\partial_x\chi(t)\partial_x\chi(t)\rangle\approx\frac{1}{2\tau_\chi^2}\langle(\nabla\chi(t))^2\rangle \tag*{(73)}

Next, we use the following approximation for the thermal average,

ℏ28me⟨(∇χ(t))2⟩≈12ΔEχ∫0∞dEχEχe−EχkBT=2ΔEχ(kBT)2(74)\frac{\hbar^2}{8m_e}\langle(\nabla\chi(t))^2\rangle\approx\frac{1}{2\Delta E_\chi}\int_0^\infty dE_\chi E_\chi e^{-\frac{E_\chi}{k_BT}}=\frac{2}{\Delta E_\chi}(k_BT)^2 \tag*{(74)}

where Eχ=ℏ28me(∇χ(t))2E_\chi=\frac{\hbar^2}{8m_e}(\nabla\chi(t))^2 is used for the energy for the vortex of χ\chi; it is the kinetic energy of mass mem_e and the velocity ℏ2me∇χ\frac{\hbar}{2m_e}\nabla\chi given in Eq. (12) with neglecting A\mathbf{A}. ΔEχ\Delta E_\chi is the typical energy fluctuation caused by the gauge fluctuation; and factor 2 takes into account the degeneracy due to the state with ∇χ\nabla\chi and that with −∇χ-\nabla\chi.

Overall, τ\tau is given by

τ≈τχΔEχkBT≈ℏkBT(75)\tau\approx\frac{\tau_{\chi}\Delta E_{\chi}}{k_{\mathrm{B}}T} \approx\frac{\hbar}{k_{\mathrm{B}}T} \tag*{(75)}

where τχΔEχ≈ℏ\tau_{\chi}\Delta E_{\chi} \approx\hbar is used in the last equality. This explains the Planckian time-scale. The above Planckian time behavior was observed in the cuprate superconductor Bi2_2Sr2_2CuO6+δ_{6+\delta} from the superconducting transition temperature Tc≈7T_{\mathrm{c}} \approx7 up to 700 K [54]. The superconductivity in the cuprate occurs in the CuO2_2 plane [55, 56]; and an estimate of the transition temperature indicates it corresponds to the stabilization of nanometer-sized loop currents in that plane [42]. The BKT transition type superconducting transition temperature behavior in the cuprate also indicates it is related to the stabilization of loop currents [57]. A computer simulation for a model for loop currents in the cuprate exhibits that the superconducting transition temperature corresponds to the temperature where the fluctuation of −ℏ2∂t(∇χ)-\frac{\hbar}{2}\partial_t(\nabla\chi) is suppressed [43].

Taking into account all the results, it is highly plausible to consider that the Planckian time behavior of the relaxation time in the cuprate superconductivity is related to the superconducting mechanism of it, where the supercurrent is generated as a collection of stable loop currents of the nanometer size in the CuO2_2 plane.

Concluding remarks

In textbooks, the Ohm’s law is explained by assuming that the effect of the battery connection is to generate electric field inside the metallic wire. During the development of the Drude theory, the force from the gradient of chemical potential was not known; thus, the forces used were the electric field force and the friction force. By considering the balance of them, the conductivity was obtained. Since the friction force causes the energy dissipation, the Joule heat was attributed to it. However, this theory is known to give an odd energy flow when the Poynting theorem is used [4, 5]. It is also notable that the steady current state appears after the drift velocity becomes constant; thus, how this velocity is maintained needs to be considered; in the present theory, it is maintained by the shift of the fluctuation center and the appearance of the induced electric field given in Eqs. (51) and (52). On the other hand, the Drude theory only considers how this drift velocity is achieved as shown in Eqs. (47) and (48). Note that the Drude theory is applicable for the optical conductivity case, where the radiation enters from outside, and absorbed by the wire.

The chemical potential employed in this work is not the thermodynamical origin, in contrast to the one introduced by Gibbs [12]. In the electron gas theory by Sommerfeld [58], the chemical potential appears in the Fermi distribution function. In the more elaborated Boltzmann equation method, the spatial variation of the chemical potential is taken into account, and the chemical potential gradient force appears. However it starts from the free electron model; on the other hand, the present method starts from the many-body wave function, thus, the many-body effect is included from the beginning.

For the explanation of the Planckian dissipation problem, ‘gauge fluctuation’, the fluctuation of −ℏ2∂t(∇χ)-\frac{\hbar}{2}\partial_t(\nabla\chi), is the key ingredient. If this fluctuation is suppressed (A — (ℏ2e∇χ)\left(\frac{\hbar}{2e}\nabla\chi\right) becomes gauge invariant; then, v\mathbf{v} that explicitly depends on A\mathbf{A} is realized. Conversely, we may say that the reason for the gauge invariance in E\mathbf{E} and B\mathbf{B} in the classical electromagnetic theory is due to the fact that it deals with cases where the ‘gauge fluctuation’ is so large that physical variables are those survive the noise from it; and E\mathbf{E} and B\mathbf{B} are such variables.

In Maxwell’s equations, A\mathbf{A} and φ\varphi do not appear, although Maxwell used them to formulate the equations [59]. Most notably, he obtained the electromagnetic wave using the vector potential [60]. Feynman stated, ‘If we take away the scaffolding he used to build it, we find that Maxwell’s beautiful edifice stands on its own’ [61]. The present work indicates the possible revision of this statement may be needed: What used to be considered as the scaffolding is actually part of the edifice; the noise from ‘gauge fluctuation’ blurs its appearance, and we have been mistakenly consider the part made of E\mathbf{E} and B\mathbf{B} exhibits the whole beauty of the edifice. If the noise is cleared, the true beauty of the edifice is revealed. It is made of A\mathbf{A} and φ\varphi with accompanying the Berry connection.

Author contributions: H.K. contributed all.

Competing interests. : The author declare that there are no competing interests in this work.

The relation between the BCS theory and the new theory using the Berry connection

For the convenience of the reader, we succinctly explain the relation between the BCS theory and the new theory that encompasses it. More detailed explanation may be found in Refs. [17, 62, 63].

The BCS superconducting state vector is given by

∣BCS(θ)⟩=∏k(uk+vkck↑†c−k↓†eiθ)∣vac⟩(76)|\mathrm{BCS}(\theta)\rangle= \prod_{\mathbf{k}}\left(u_{\mathbf{k}}+v_{\mathbf{k}}c_{\mathbf{k}\uparrow}^{\dagger}c_{-\mathbf{k}\downarrow}^{\dagger}e^{i\theta}\right)|\mathrm{vac}\rangle \tag*{(76)}

where real parameters uku_{\mathbf{k}} and vkv_{\mathbf{k}} satisfy uk2+vk2=1u_{\mathbf{k}}^{2}+v_{\mathbf{k}}^{2}=1. It is a linear combination of different particle number states, thus, breaks the global U(1)U(1) gauge invariance. The spatial variation of θ\theta (which is absent in Eq. (76) but appears later in Eq. (83)) is the Nambu-Goldstone mode and the vector potential A\mathbf{A} appears in the theory as the following gauge invariant combination

A+ℏ2e∇θ(77)\mathbf{A}+\frac{\hbar}{2e}\nabla\theta \tag*{(77)}

The BCS theory is systematically described using the following Bogoliubov operators [31] given by

γk↑BCS=uke−i2θck↑−vkei2θc−k↓†γ−k↓BCS=uke−i2θc−k↓+vkei2θck↑†(78)\begin{aligned} \gamma^{\mathrm{BCS}}_{k\uparrow} &= u_k e^{-\frac{i}{2}\theta}c_{k\uparrow}-v_k e^{\frac{i}{2}\theta}c^\dagger_{-k\downarrow}\\ \gamma^{\mathrm{BCS}}_{-k\downarrow} &= u_k e^{-\frac{i}{2}\theta}c_{-k\downarrow}+v_k e^{\frac{i}{2}\theta}c^\dagger_{k\uparrow} \tag*{(78)} \end{aligned}

The superconducting ground state is defined as the ground state for the Bogoliubov excitations

γk↑BCS∣BCS⟩=0,γ−k↓BCS∣BCS⟩=0(79)\gamma^{\mathrm{BCS}}_{k\uparrow}|\mathrm{BCS}\rangle=0,\qquad\gamma^{\mathrm{BCS}}_{-k\downarrow}|\mathrm{BCS}\rangle=0 \tag*{(79)}

The difference of the field operators in the normal and superconducting states is crucial. The normal metallic state is assumed to be well-described by the free electrons with the effective mass m∗m^*. Then, the electron field operators are given by

Ψ^σnorm(r)=1V∑keik⋅rckσ(80)\hat{\Psi}^{\mathrm{norm}}_{\sigma}(\mathbf{r})=\frac{1}{\sqrt{V}}\sum_{\mathbf{k}}e^{i\mathbf{k}\cdot\mathbf{r}}c_{\mathbf{k}\sigma} \tag*{(80)}

On the other hand, in the superconducting state, the electron field operators are expressed using the Bogoliubov operators as follows

Ψ^↑BCS(r)=∑kei2θ[γk↑BCSuk(r)−(γBCS)−k↓†vk(r)]Ψ^↓BCS(r)=∑kei2θ[γk↓BCSuk(r)+(γBCS)−k↑†vk(r)](81)\begin{aligned} \hat{\Psi}^{\mathrm{BCS}}_{\uparrow}(\mathbf{r})&=\sum_{\mathbf{k}}e^{\frac{i}{2}\theta}\left[\gamma^{\mathrm{BCS}}_{k\uparrow}u_k(\mathbf{r})-\left(\gamma^{\mathrm{BCS}}\right)^\dagger_{-k\downarrow}v_k(\mathbf{r})\right]\\ \hat{\Psi}^{\mathrm{BCS}}_{\downarrow}(\mathbf{r})&=\sum_{\mathbf{k}}e^{\frac{i}{2}\theta}\left[\gamma^{\mathrm{BCS}}_{k\downarrow}u_k(\mathbf{r})+\left(\gamma^{\mathrm{BCS}}\right)^\dagger_{-k\uparrow}v_k(\mathbf{r})\right] \tag*{(81)} \end{aligned}

where the following coordinate dependent functions are introduced

uk(r)=1Veik⋅ruk,vk(r)=1Veik⋅rvk(82)u_k(\mathbf{r})=\frac{1}{\sqrt{V}}e^{i\mathbf{k}\cdot\mathbf{r}}u_k,\qquad v_k(\mathbf{r})=\frac{1}{\sqrt{V}}e^{i\mathbf{k}\cdot\mathbf{r}}v_k \tag*{(82)}

For considering the case where the coordinate dependent functions are different from plane waves, we use the label nn in place of the wave number k\mathbf{k}. Then, the field operators become

Ψ^↑BCS(r)=∑nei2θ(r)[γn↑BCSun(r)−(γBCS)n↓†vn∗(r)]Ψ^↓BCS(r)=∑nei2θ(r)[γn↓BCSun(r)+(γBCS)n↑†vn∗(r)](83)\begin{aligned} \hat{\Psi}^{\mathrm{BCS}}_{\uparrow}(\mathbf{r})&=\sum_n e^{\frac{i}{2}\theta(\mathbf{r})}\left[\gamma^{\mathrm{BCS}}_{n\uparrow}u_n(\mathbf{r})-\left(\gamma^{\mathrm{BCS}}\right)^\dagger_{n\downarrow}v_n^*(\mathbf{r})\right]\\ \hat{\Psi}^{\mathrm{BCS}}_{\downarrow}(\mathbf{r})&=\sum_n e^{\frac{i}{2}\theta(\mathbf{r})}\left[\gamma^{\mathrm{BCS}}_{n\downarrow}u_n(\mathbf{r})+\left(\gamma^{\mathrm{BCS}}\right)^\dagger_{n\uparrow}v_n^*(\mathbf{r})\right] \tag*{(83)} \end{aligned}

where the coordinate dependent θ\theta, θ(r)\theta(\mathbf{r}), is also introduced. With this θ(r)\theta(\mathbf{r}), the combination in Eq. (77) becomes gauge invariant. The formalism that uses the above field operators is the essence of the Bogoliubov-de Gennes formalism [64, 65]. At this point, we may view the superconducting state in the standard theory is the theory that goes beyond the Schrödinger’s wave mechanics by breaking the global U(1)U(1) gauge invariance; and the phase factor ei2θ(r)e^{\frac{i}{2}\theta(\mathbf{r})} plays the important role to produce the supercurrent [66, 67].

The new theory is constructed by comparing Eqs. (16) and (83); we notice that ei2θ(r)e^{\frac{i}{2}\theta(\mathbf{r})} may describe the phase factor with the Berry connection. Then, the following correspondence may be possible

θ→−χ^(84)\theta\rightarrow-\hat{\chi} \tag*{(84)}

where χ^\hat{\chi} is the operator version of χ\chi. Here, we use the operator version of χ\chi, χ^\hat{\chi}, is used so that the change of the number of electrons participate in the collective mode described by χ\chi (we call it, the ‘χ\chi mode’) is possible.

The χ\chi mode produces supercurrent in a similar manner as θ(r)\theta(\mathbf{r}) does. When it is quantized its canonical conjugate variable is the number of particles participating in it as shown below: In order to find the conjugate variable of χ\chi, we use the following Lagrangian

L=⟨Ψ∣iℏ∂∂t−H∣Ψ⟩=∫d3rℏ2χ˙ρχ+⋯(85)\begin{aligned} L &= \langle\Psi|i\hbar\frac{\partial}{\partial t}-H|\Psi\rangle\\ &= \int d^{3}r\frac{\hbar}{2}\dot{\chi}\rho_{\chi}+\cdots \tag*{(85)} \end{aligned}

where ρχ\rho_{\chi} is the number density of electrons participating in the χ\chi mode, and Ψ\Psi is the state vector corresponding to the wave function in Eq. (16). This indicates that the conjugate field of χ\chi, πχ\pi_{\chi}, is given by

πχ=δLδχ˙=ℏ2ρχ(86)\pi_{\chi}=\frac{\delta L}{\delta\dot{\chi}}=\frac{\hbar}{2}\rho_{\chi} \tag*{(86)}

Then, the canonical quantization condition

[χ^(r,t),π^χ(r′,t)]=iℏδ(r−r′)(87)[\hat{\chi}(\mathbf{r},t),\hat{\pi}_{\chi}(\mathbf{r}',t)]=i\hbar\delta(\mathbf{r}-\mathbf{r}') \tag*{(87)}

yields

[χ^(r,t),ρ^χ(r′,t)]=2iδ(r−r′)(88)[\hat{\chi}(\mathbf{r},t),\hat{\rho}_{\chi}(\mathbf{r}',t)]=2i\delta(\mathbf{r}-\mathbf{r}') \tag*{(88)}

for the quantization condition for χ^\hat{\chi}, where ρ^χ\hat{\rho}_{\chi} is the operator for the number density of electrons that participate in the χ\chi mode. Using the commutation relation in Eq. (88), it can be shown that ei2χ^e^{\frac{i}{2}\hat{\chi}} is the number changing operators that increases the number of electrons participating the collective motion by one, and e−i2χ^e^{-\frac{i}{2}\hat{\chi}} decreases the number by one.

By replacing θ(r)\theta(\mathbf{r}) in Eq. (83) by −χ^-\hat{\chi}, the field operators for the new theory are given by

Ψ^↑(r)=∑ne−i2χ^(r)[γn↑un(r)−γn↓†vn∗(r)]\hat{\Psi}_{\uparrow}(\mathbf{r})=\sum_n e^{-\frac{i}{2}\hat{\chi}(\mathbf{r})}\left[\gamma_{n\uparrow}u_n(\mathbf{r})-\gamma^\dagger_{n\downarrow}v_n^*(\mathbf{r})\right]
Ψ^↓(r)=∑ne−i2χ^(r)[γn↓un(r)+γn↑†vn∗(r)](89)\hat{\Psi}_{\downarrow}(\mathbf{r})=\sum_n e^{-\frac{i}{2}\hat{\chi}(\mathbf{r})}\left[\gamma_{n\downarrow}u_n(\mathbf{r})+\gamma^\dagger_{n\uparrow}v_n^*(\mathbf{r})\right] \tag*{(89)}

The important point to note is that the above Bogoliubov operators conserve particle numbers; thus we call them the particle number conserving Bogoliubov operators (PNC-BOs). The PNC-BOs cause the fluctuation of the number of electrons in the collective mode.

The PNC-BOs satisfy

γnσ∣Gnd(N)⟩=0,⟨Gnd(N)∣γnσ†=0(90)\gamma_{n\sigma}|{\rm Gnd}(N)\rangle=0,\qquad\langle{\rm Gnd}(N)|\gamma^\dagger_{n\sigma}=0 \tag*{(90)}

where NN is the total number of particles (see Eq. (79) for the BCS case). They contain an additional operator ei2χ^(r)e^{\frac{i}{2}\hat{\chi}(\mathbf{r})}. We take the ground state to be the eigenstate of it that satisfies

ei2χ^(r)∣Gnd(N)⟩=ei2χ(r)∣Gnd(N+1)⟩(91)e^{\frac{i}{2}\hat{\chi}(\mathbf{r})}|{\rm Gnd}(N)\rangle=e^{\frac{i}{2}\chi(\mathbf{r})}|{\rm Gnd}(N+1)\rangle \tag*{(91)}

At this point, we may view the superconducting state in the new theory as the theory that goes beyond the Schrödinger’s quantum mechanics by including the quantized version of the neglected U(1)U(1) phase by Dirac [68]. As is explained in Ref. [17], the usual momentum operator adopted by Schrödinger can have an emergent gauge field contribution, which is given as the Berry connection in Eq. (5). The new superconductivity theory includes it in the quantized form given in Eq. (88).

Using the PNC-BO, we can construct the particle number conserving Bogoliubov-de Gennes equations (PNC-BdG equations) as follows: Let us consider the following electronic Hamiltonian

He=∑σ∫d3r Ψ^σ†(r)h(r)Ψ^σ(r)−12∑σ,σ′∫d3r d3r′ Veff(r,r′)Ψ^σ†(r)Ψ^σ′†(r′)Ψ^σ′(r′)Ψ^σ(r)(92)H_e=\sum_{\sigma}\int d^3r\,\hat{\Psi}^{\dagger}_{\sigma}(\mathbf{r})h(\mathbf{r})\hat{\Psi}_{\sigma}(\mathbf{r})-\frac{1}{2}\sum_{\sigma,\sigma'}\int d^3r\,d^3r'\,V_{\mathrm{eff}}(\mathbf{r},\mathbf{r}')\hat{\Psi}^{\dagger}_{\sigma}(\mathbf{r})\hat{\Psi}^{\dagger}_{\sigma'}(\mathbf{r}')\hat{\Psi}_{\sigma'}(\mathbf{r}')\hat{\Psi}_{\sigma}(\mathbf{r}) \tag*{(92)}

where h(r)h(\mathbf{r}) is the single-particle Hamiltonian given by

h(r)=12me(ℏi∇+ecA)2+U(r)−μ(93)h(\mathbf{r})=\frac{1}{2m_e}\left(\frac{\hbar}{i}\nabla+\frac{e}{c}\mathbf{A}\right)^2+U(\mathbf{r})-\mu \tag*{(93)}

and −Veff-V_{\mathrm{eff}} is the effective interaction between electrons.

We perform the mean field approximation on HeH_e. The result is

HeMF=∑σ∫d3r Ψ^σ†(r)h(r)Ψ^σ(r)+∫d3rd3r′[Δ(r,r′)Ψ^↑†(r)Ψ^↓†(r′)e−i2(χ^(r)+χ^(r′))+H.c.]+∫d3rd3r′∣Δ(r,r′)∣2Veff(r,r′)(94)H_e^{\mathrm{MF}}=\sum_{\sigma}\int d^3r\,\hat{\Psi}^{\dagger}_{\sigma}(\mathbf{r})h(\mathbf{r})\hat{\Psi}_{\sigma}(\mathbf{r})+\int d^3r d^3r'\left[\Delta(\mathbf{r},\mathbf{r}')\hat{\Psi}^{\dagger}_{\uparrow}(\mathbf{r})\hat{\Psi}^{\dagger}_{\downarrow}(\mathbf{r}')e^{-\frac{i}{2}(\hat{\chi}(\mathbf{r})+\hat{\chi}(\mathbf{r}'))}+\mathrm{H.c.}\right] +\int d^3r d^3r'\frac{|\Delta(\mathbf{r},\mathbf{r}')|^2}{V_{\mathrm{eff}}(\mathbf{r},\mathbf{r}')} \tag*{(94)}

where the gap function Δ(r,r′)\Delta(\mathbf{r},\mathbf{r}') is defined as

Δ(r,r′)=Veff(r,r′)⟨ei2(χ^(r)+χ^(r′))Ψ^↑(r)Ψ^↓(r′)⟩(95)\Delta(\mathbf{r},\mathbf{r}')=V_{\mathrm{eff}}(\mathbf{r},\mathbf{r}')\left\langle e^{\frac{i}{2}(\hat{\chi}(\mathbf{r})+\hat{\chi}(\mathbf{r}'))}\hat{\Psi}_{\uparrow}(\mathbf{r})\hat{\Psi}_{\downarrow}(\mathbf{r}')\right\rangle \tag*{(95)}

Due to the factor ei2(χ^(r)+χ^(r′))e^{\frac{i}{2}(\hat{\chi}(\mathbf{r})+\hat{\chi}(\mathbf{r}'))} the expectation value in Δ\Delta can be calculated using the particle number fixed state.

Using commutation relations for Ψ^σ†(r)\hat{\Psi}^{\dagger}_{\sigma}(\mathbf{r}) and Ψ^σ(r)\hat{\Psi}_{\sigma}(\mathbf{r}),

{Ψ^σ(r),Ψ^σ′†(r′)}=δσσ′δ(r−r′){Ψ^σ(r),Ψ^σ′(r′)}=0{Ψ^σ†(r),Ψ^σ′†(r′)}=0(96)\begin{aligned} \{\hat{\Psi}_{\sigma}(\mathbf{r}),\hat{\Psi}^{\dagger}_{\sigma'}(\mathbf{r}')\}&=\delta_{\sigma\sigma'}\delta(\mathbf{r}-\mathbf{r}')\\ \{\hat{\Psi}_{\sigma}(\mathbf{r}),\hat{\Psi}_{\sigma'}(\mathbf{r}')\}&=0\\ \{\hat{\Psi}^{\dagger}_{\sigma}(\mathbf{r}),\hat{\Psi}^{\dagger}_{\sigma'}(\mathbf{r}')\}&=0 \tag*{(96)} \end{aligned}

the following relations are obtained

[Ψ^↑(r),HMF]=h(r)Ψ^↑(r)+∫d3r′Δ(r,r′)Ψ^↓†(r′)e−i2(χ^(r)+χ^(r′))[\hat{\Psi}_{\uparrow}(\mathbf{r}),H_{\mathrm{MF}}]=h(\mathbf{r})\hat{\Psi}_{\uparrow}(\mathbf{r})+\int d^3r'\Delta(\mathbf{r},\mathbf{r}')\hat{\Psi}^{\dagger}_{\downarrow}(\mathbf{r}')e^{-\frac{i}{2}(\hat{\chi}(\mathbf{r})+\hat{\chi}(\mathbf{r}'))}
[Ψ^↓(r),HMF]=h(r)Ψ^↓(r)−∫d3r′Δ(r,r′)Ψ^↑†(r′)e−i2(χ^(r)+χ^(r′))(97)[\hat{\Psi}_{\downarrow}(\mathbf{r}),H_{\mathrm{MF}}]=h(\mathbf{r})\hat{\Psi}_{\downarrow}(\mathbf{r})-\int d^3r'\Delta(\mathbf{r},\mathbf{r}')\hat{\Psi}^{\dagger}_{\uparrow}(\mathbf{r}')e^{-\frac{i}{2}(\hat{\chi}(\mathbf{r})+\hat{\chi}(\mathbf{r}'))} \tag*{(97)}

The PNC-BOs γnσ\gamma_{n\sigma} and γnσ†\gamma^{\dagger}_{n\sigma} are the fermion operators chosen to satisfy the following relations

[HMF,γnσ]=−ϵnγnσ,[HMF,γnσ†]=ϵnγnσ†(98)[H_{\mathrm{MF}},\gamma_{n\sigma}]=-\epsilon_n\gamma_{n\sigma},\qquad[H_{\mathrm{MF}},\gamma^{\dagger}_{n\sigma}]=\epsilon_n\gamma^{\dagger}_{n\sigma} \tag*{(98)}

Then, HMFH_{\mathrm{MF}} is diagonalized as

HMF=Eg+∑n,σ′ϵnγnσ†γnσ(99)H_{\mathrm{MF}}=E_g+\sum_{n,\sigma}'\epsilon_n\gamma^{\dagger}_{n\sigma}\gamma_{n\sigma} \tag*{(99)}

where EgE_g is the ground state energy, and ‘∑′\sum' indicates the sum is taken over ϵn≥0\epsilon_n\geq0 states.

From Eqs. (A22), (A23), and (A24), we obtain the following system of equations

ϵnun(r)=ei2χ^(r)h(r)e−i2χ^(r)un(r)+ei2χ^(r)∫d3r′Δ(r,r′)e−i2χ^(r′)vn(r′)\epsilon_n u_n(\mathbf{r})=e^{\frac{i}{2}\hat{\chi}(\mathbf{r})}h(\mathbf{r})e^{-\frac{i}{2}\hat{\chi}(\mathbf{r})}u_n(\mathbf{r})+e^{\frac{i}{2}\hat{\chi}(\mathbf{r})}\int d^3r'\Delta(\mathbf{r},\mathbf{r}')e^{-\frac{i}{2}\hat{\chi}(\mathbf{r}')}v_n(\mathbf{r}')
ϵnvn∗(r)=−ei2χ(r)h(r)e−i2χ(r)vn∗(r)+ei2χ(r)∫d3r′ Δ(r,r′)e−i2χ(r′)un∗(r′)\epsilon_n v_n^*(\mathbf r)=-e^{\frac{i}{2}\chi(\mathbf r)}h(\mathbf r)e^{-\frac{i}{2}\chi(\mathbf r)}v_n^*(\mathbf r)+e^{\frac{i}{2}\chi(\mathbf r)}\int d^3r'\,\Delta(\mathbf r,\mathbf r')e^{-\frac{i}{2}\chi(\mathbf r')}u_n^*(\mathbf r')

We take the expectation value of the above equations for the state ∣Gnd(N)⟩|\mathrm{Gnd}(N)\rangle. Using the relation in Eq. (91), the above are cast into the following,

ϵnun(r)=hˉ(r)un(r)+∫d3r′ Δ(r,r′)vn(r′)ϵnvn(r)=−hˉ∗(r)vn(r)+∫d3r′ Δ∗(r,r′)un(r′)(100)\begin{aligned} \epsilon_n u_n(\mathbf r)&=\bar{h}(\mathbf r)u_n(\mathbf r)+\int d^3r'\,\Delta(\mathbf r,\mathbf r')v_n(\mathbf r')\\ \epsilon_n v_n(\mathbf r)&=-\bar{h}^*(\mathbf r)v_n(\mathbf r)+\int d^3r'\,\Delta^*(\mathbf r,\mathbf r')u_n(\mathbf r') \tag*{(100)} \end{aligned}

where the single particle Hamiltonian hˉ\bar{h} is

hˉ(r)=12me(ℏi∇+ecA−ℏ2∇χ)2+U(r)−μ(101)\bar{h}(\mathbf r)=\frac{1}{2m_e}\left(\frac{\hbar}{i}\nabla+\frac{e}{c}\mathbf A-\frac{\hbar}{2}\nabla\chi\right)^2+U(\mathbf r)-\mu \tag*{(101)}

the pair potential Δ(r,r′)\Delta(\mathbf r,\mathbf r') is

Δ(r,r′)=Veff∑n′[un(r)vn∗(r′)(1−f(ϵn))−un(r′)vn∗(r)f(ϵn)](102)\Delta(\mathbf r,\mathbf r')=V_{\mathrm{eff}}\sum_n'\left[u_n(\mathbf r)v_n^*(\mathbf r')(1-f(\epsilon_n))-u_n(\mathbf r')v_n^*(\mathbf r)f(\epsilon_n)\right] \tag*{(102)}

and f(ϵn)f(\epsilon_n) is the Fermi function (T→0T\to0 limit should be considered for the ground state). The number density is given by

ρ(r)=∑n′[∣un(r)∣2f(ϵn)+∣vn(r)∣2(1−f(ϵn))](103)\rho(\mathbf r)=\sum_n'\left[|u_n(\mathbf r)|^2f(\epsilon_n)+|v_n(\mathbf r)|^2(1-f(\epsilon_n))\right] \tag*{(103)}

They are the PNC-BdG equations [69]. The gauge potential in the single particle Hamiltonian hˉ(r)\bar{h}(\mathbf r) is the effective one given by

Aeff=A−ℏ2e∇χ(104)\mathbf A^{\mathrm{eff}}=\mathbf A-\frac{\hbar}{2e}\nabla\chi \tag*{(104)}

This combination corresponds to Eq. (77), and appeared in the velocity field in Eq. (12).

Now the total energy is a functional of A−ℏ2e∇χ\mathbf A-\frac{\hbar}{2e}\nabla\chi, its time-component partner φ+ℏ2e∂tχ\varphi+\frac{\hbar}{2e}\partial_t\chi, and ρ\rho. We may write it in the following form

H=∫d3r[H(A−ℏ2e∇χ,φ+ℏ2e∂tχ,ρ)−eρφ](105)H=\int d^3r\left[\mathcal H\left(\mathbf A-\frac{\hbar}{2e}\nabla\chi,\varphi+\frac{\hbar}{2e}\partial_t\chi,\rho\right)-e\rho\varphi\right] \tag*{(105)}

Note that the last term is the contribution from the electrostatic potential, and the number density ρ\rho is the sum of ρχ\rho_\chi and the rest. The above functional should be compared with the one proposed by Oliveira, Gross, and Kohn [70], which uses Δ\Delta as the extra parameter to describe the superconducting state. In the present case, the superconducting state is characterized by (∇χ,∂tχ)(\nabla\chi,\partial_t\chi); they originate from the neglected U(1)U(1) phase by Dirac [17], which upsets the fundamental assumption of the density functional theory [71] since the use of the operator form of the momentum p=−iℏ∇\mathbf{p}=-i\hbar\nabla is violated. In other words, the present formalism indicates that the superconducting state is related to the violation of using p=−iℏ∇\mathbf{p}=-i\hbar\nabla. Note that the present formalism agrees with the claim that the fundamental physics occurring in superconductivity is the gauge symmetry breaking accompanied by the appearance of the Nambu-Goldstone mode [66, 67].

Employing the classical counterpart of the commutation relation in Eq. (88), i.e., taking ℏ2χ\frac{\hbar}{2}\chi and ρχ\rho_{\chi} as canonical conjugate classical variables, we obtain the following Hamilton equations

ℏ2∂tχ=∂H∂ρχ−eφ(106)\frac{\hbar}{2}\partial_t\chi=\frac{\partial\mathcal{H}}{\partial\rho_{\chi}}-e\varphi \tag*{(106)}
∂tρχ=∇⋅∂H∂(ℏ2∇χ)=−1e∇⋅∂H∂A=1e∇⋅j(107)\partial_t\rho_{\chi}=\nabla\cdot\frac{\partial\mathcal{H}}{\partial\left(\frac{\hbar}{2}\nabla\chi\right)} =-\frac{1}{e}\nabla\cdot\frac{\partial\mathcal{H}}{\partial\mathbf{A}} =\frac{1}{e}\nabla\cdot\mathbf{j} \tag*{(107)}

The equation (A31) corresponds to Eq. (13) with μ=∂H∂ρχ\mu=\frac{\partial\mathcal{H}}{\partial\rho_{\chi}}. The equation (A32) describes the local charge conservation with j=−eρχv\mathbf{j}=-e\rho_{\chi}\mathbf{v}, where v\mathbf{v} is given by Eq. (12).

References

  1. [1]Ohm, G.S.: Die Galvanische Kette, Mathematisch Bearbeitet. T. H. Riemann, Berlin (1827)
  2. [2]Maxwell, J.C. (ed.): The Electrical Researches of the Honourable Henry Cavendish, F.R.S., Written Between 1771 and 1781. Cambridge University Press, London (1879)
  3. [3]Joule, J.P.: On the heat evolved by metallic conductors of electricity, and in the cells of a battery during electrolysis. Philosophical Magazine 19, 260 (1841)DOI
  4. [4]Feynman, R.P., Leighton, R.B., Sands, M.: The Feynman Lectures on Physics, vol. II. Addison-Wesley, Reading (1963). Chap. 27-5DOI
  5. [5]Mörée, G., Leijon, M.: Comparison of poynting’s vector and the power flow used in electrical engineering. AIP Advances 12(8), 085219 (2022) https://doi.org/10.1063/5.0101339 https://pubs.aip.org/aip/adv/article-pdf/doi/10.1063/5.0101339/16468542/085219_1_online.pdf
  6. [6]Poynting, J.H.: On the transfer of energy in the electromagnetic field. Philosophical Transactions of the Royal Society of London 175, 343–361 (1884)DOI
  7. [7]Drude, P.: Zur elektronentheorie der metalle. Annalen der Physik 306, 566–613 (1900)
  8. [8]Lorentz, H.A.: Le mouvement des électrons dans les métaux. Ach. néerl. 10, 336 (1905)
  9. [9]Lorentz, H.A.: The Theory of Electrons and Its Applications to the Phenomena of Light and Radiant Heat, 2nd edn. Dover, New York (1952)
  10. [10]Bohr, N.: Studier over Metallernes Elektrontheori. Thaning & Appel, Kopenhagen (1911)
  11. [11]Koizumi, H.: Constrained hamiltonian dynamics for electrons in magnetic field and additional forces besides the lorentz force acting on electrons. J. Phys. A: Math. Theor. 57, 435304 (2024)
  12. [12]Gibbs, J.W.: On the equilibrium of heterogeneous substances. first part. Transactions of the Connecticut Academy of Arts and Sciences 3, 108–248 (1876)
  13. [13]Landauer, R.: Spatial variation of currents and fields due to localized scatterers in metallic conduction. IBM J. Res. Dev. 1, 223 (1957)
  14. [14]Büttiker, M., Imry, Y., Landauer, R., Pinhas, S.: Generalized many-channel conductance formula with application to small rings. Phys. Rev. B 31, 6207–6215 (1985) https://doi.org/10.1103/PhysRevB.31.6207
  15. [15]Datta, S.: Electronic Transports in Mesoscopic Systems. Cambridge Univ. Press, Cambridge (1995)
  16. [16]Koizumi, H.: Schrödinger representation of quantum mechanics, berry connection, and superconductivity. Physics Letters A 450, 128367 (2022) https://doi.org/10.1016/j.physleta.2022.128367
  17. [17]Koizumi, H.: Neglected u(1) phase in the schroedinger representation of quantum mechanics and particle number conserving formalisms for superconductivity. Journal of Physics A: Mathematical and Theoretical 56, 455303 (2023)arxiv.org/abs/2211.08759
  18. [18]Sachdev, S., Ye, J.: Universal quantum-critical dynamics of two-dimensional antiferromagnets. Phys. Rev. Lett. 69, 2411–2414 (1992) https://doi.org/10.1103/PhysRevLett.69.2411
  19. [19]Zaanen, J.: Planckian dissipation, minimal viscosity and the transport in cuprate strange metals. SciPost Phys. 6, 061 (2019) https://doi.org/10.21468/SciPostPhys.6.5.061
  20. [20]Hartnoll, S.A., Mackenzie, A.P.: Colloquium: Planckian dissipation in metals. Rev. Mod. Phys. 94, 041002 (2022) https://doi.org/10.1103/RevModPhys.94.041002
  21. [21]Bardeen, J., Cooper, L.N., Schrieffer, J.R.: Theory of superconductivity. Phys. Rev. 108, 1175 (1957)
  22. [22]Josephson, B.D.: Possible new effects in superconductive tunneling. Phys. Lett. 1, 251 (1962)DOI
  23. [23]Feynman, R.P., Leighton, R.B., Sands, M.: The Feynman Lectures on Physics, vol. III. Addison-Wesley, Reading (1963). Chap. 21-9
  24. [24]Schrieffer, J.R.: Theory Of Superconductivity. Benjamin-Cummings Publishing Company, San Francisco (1983). Sec. 8-8
  25. [25]Murani, A., Bourlet, N., Sueur, H., Portier, F., Altimiras, C., Esteve, D., Grabert, H., Stockburger, J., Ankerhold, J., Joyez, P.: Absence of a dissipative quantum phase transition in josephson junctions. Phys. Rev. X 10, 021003 (2020) https://doi.org/10.1103/PhysRevX.10.021003
  26. [26]Aumentado, J., Catelani, G., Serniak, K.: Quasiparticle poisoning in superconducting quantum computers. Physics Today 76(8), 34–39 (2023)
  27. [27]Serniak, K.: Nonequilibrium quasiparticles in superconducting qubits. PhD thesis, Yale University (2019)
  28. [28]Koizumi, H., Ishikawa, A.: Berry connection from many-body wave functions and superconductivity: Calculations by the particle number conserving bogoliubov-de gennes equations. Journal of Superconductivity and Novel Magnetism 34, 2795 (2021) https://doi.org/10.1007/s10948-021-05991-y
  29. [29]Anderson, P.W.: Considerations on the flow of superfluid helium. Rev. Mod. Phys. 38, 298 (1966)
  30. [30]Maxwell, J.C.: A Treatise on Electricity and Magnetism, 3rd edn., p. 297. Dover, New York (1954)
  31. [31]Bogoliubov, N.N.: A new method in the theory of superconductivity I. Soviet Physics JETP 7, 41 (1958)
  32. [32]Valatin, J.G. Nuovo Cimento 7, 843 (1958)
  33. [33]Peierls, R.: Spontaneously broken symmetries. J. Phys. A 24, 5273 (1991)DOI
  34. [34]Peierls, R.: Broken symmetries. Contemporary Phys. 33, 221 (1992)
  35. [35]Bohm, D.: Note on a theorem of Bloch concerning possible causes of superconductivity. Phys. Rev 75, 502 (1949)DOI
  36. [36]Bohm, D., Pines, D.: A collective description of electron interactions: Iii. coulomb interactions in a degenerate electron gas. Phys. Rev. 92, 609–625 (1953) https://doi.org/10.1103/PhysRev.92.609
  37. [37]Anderson, P.W.: Coherent excited states in the theory of superconductivity: Gauge invariance and the meissner effect. Phys. Rev. 110, 827 (1958)DOI
  38. [38]Anderson, P.W.: Random-phase approximation in the theory of superconductivity. Phys. Rev. 112, 1900 (1958)DOI
  39. [39]Fetter, A., Walecka, J.F.: Quantum Theory of Many-Particle Systems, pp. 182–183. McGraw-Hill, San Francisco (1971)
  40. [40]Pines, D.: Elementary Excitations in Solids, p. 96. CRC Press, Boca Raton, Florida (1999)
  41. [41]Bednorz, J.G., Müller, K.A.: Possible high t_c superconductivity in the Ba-La-Cu-O system. Z. Phys. B 64, 189 (1986)DOI
  42. [42]Emery, V.J., Kivelson, S.A.: Importance of phase fluctuation in superconductors with small superfluid density. Nature 374, 434 (1995)DOI
  43. [43]Okazaki, A., Wakaura, H., Koizumi, H., Ghantous, M.A., Tachiki, M.: Superconducting transition temperature of the hole-doped cuprate as the stabilization temperature of supercurrent loops generated by spin-twisting itinerant motion of electrons. J. Supercond. Nov. Magn. 28, 3221–3233 (2015)DOI
  44. [44]Hirsch, J.E.: The London moment: what a rotating superconductor reveals about superconductivity. Physica Scripta 89(1), 015806 (2013) https://doi.org/10.1088/0031-8949/89/01/015806
  45. [45]Hirsch, J.E.: The lorentz force and superconductivity. Phys. Lett. A 315, 474–479 (2003)DOI
  46. [46]Hirsch, J.E.: Momentum of superconducting electrons and the explanation of the Meissner effect. Phys Rev. B 95, 014503 (2017)DOI
  47. [47]Koizumi, H.: Supercurrent and electromotive force generations by the berry connection from many-body wave functions. Journal of Physics A: Mathematical and Theoretical 56(18), 185301 (2023) https://doi.org/10.1088/1751-8121/acc858
  48. [48]Kampen, N.G.V.: Stochastic Processes in Physics and Chemistry, 3rd edn. Elsevier, Amsterdam (2007)
  49. [49]Gardiner, C., Zoller, P.: Quantum Noise, 3rd edn. Springer, New York (2004)
  50. [50]Feynman, R.P., Leighton, R.B., Sands, M.: The Feynman Lectures on Physics, vol. II. Addison-Wesley, Reading (1963). Sec. 21-3DOI
  51. [51]Maxwell, J.C.: A Treatise on Electricity and Magnetism, 3rd edn., A617. Dover, New York (1954)
  52. [52]Shapiro, S.: Josephson currents in superconducting tunneling: the effect of microwaves and other observations. Phys. Rev. Lett. 11, 80 (1963)DOI
  53. [53]Koizumi, H.: Superconductivity by berry connection from many-body wave functions: Revisit to Andreev-Saint-James reflection and Josephson effect. Journal of Superconductivity and Novel Magnetism 34(8), 2017–2029 (2021) https://doi.org/10.1007/s10948-021-05905-y
  54. [54]Martin, S., Fiory, A.T., Fleming, R.M., Schneemeyer, L.F., Waszczak, J.V.: Normal-state transport properties of bi₂₊ₓsr₂₋ᵧcuo₆₊δ crystals. Phys. Rev. B 41, 846–849 (1990) https://doi.org/10.1103/PhysRevB.41.846
  55. [55]Jiang, D., Hu, T., You, L., Li, Q., Li, A., Wang, H., Mu, G., Chen, Z., Zhang, H., Yu, G., Zhu, J., Sun, Q., Lin, C., Xiao, H., Xie, X., Jiang, M.: High-Tc superconductivity in ultrathin Bi2Sr2CaCu2O(8+x) down to half-unit-cell thickness by protection with graphene. Nat Commun 5, 5708 (2014) https://doi.org/10.1038/ncomms6708
  56. [56]Yu, Y., Ma, L., Cai, P., Zhong, R., Ye, C., Shen, J., Gu, G.D., Chen, X.H., Zhang, Y.: High-temperature superconductivity in monolayer Bi₂Sr₂CaCu₂O₈₊δ. Nature 575(7781), 156–163 (2019) https://doi.org/10.1038/s41586-019-1718-x
  57. [57]Zhang, L., Kang, C., Liu, C., Wang, K., Zhang, W.: Two-dimensional superconducting nature of Bi₂Sr₂CaCu₂O₈₊δ thin films revealed by crystal structure transition. RSC Adv. 13, 25797–25803 (2023) https://doi.org/10.1039/D3RA02701E
  58. [58]Sommerfeld, A.: Zur elektronentheorie der metalle auf grund der fermischen statistik. Zeitschrift für Physik 47(1), 1–32 (1928) https://doi.org/10.1007/BF01391052
  59. [59]Maxwell, J.C.: A Treatise on Electricity and Magnetism, 3rd edn., A618. Dover, New York (1954)
  60. [60]Maxwell, J.C.: A Treatise on Electricity and Magnetism, 3rd edn., A784. Dover, New York (1954)
  61. [61]Feynman, R.P., Leighton, R.B., Sands, M.: The Feynman Lectures on Physics, vol. II. Addison-Wesley, Reading (1963). Sec. 18-1DOI
  62. [62]Koizumi, H.: London Moment, London’s Superpotential, Nambu-Goldstone Mode, and Berry Connection from Many-Body Wave Functions. Journal of Superconductivity and Novel Magnetism 34(5), 1361–1370 (2021) https://doi.org/10.1007/s10948-021-05827-9
  63. [63]Taya, H., Takatsu, Y., Koizumi, H.: Toward the detection of spin-vortex-induced loop currents in a single bilayer Bi₂Sr₂CaCu₂O₈₊δ thin film and their possible use as qubits: Model calculations for three nano-island architecture. to appear in J. Supercond. Nov. Magn.
  64. [64]Gennes, P.G.: Superconductivity of Metals and Alloys. W. A. Benjamin, Inc., ??? (1966)
  65. [65]Zhu, J.-X.: Bogoliubov-de Gennes Method and Its Applications. Springer, ??? (2016)DOI
  66. [66]Anderson, P.W.: Basic Notions of Condensed Matter Physics. The Benjamin/Cummings Publishing Company, Inc., San Francisco (1984)
  67. [67]Weinberg, S.: Superconductivity for particular theorists. Progress of Theoretical Physics Supplement 86, 43–53 (1986) https://doi.org/10.1143/PTPS.86.43
  68. [68]Dirac, P.A.M.: Principles of Quantum Mechanics, 4th edn. Oxford Univ. Press, Oxford (1958)
  69. [69]Koizumi, H.: Explanation of superfluidity using the Berry connection for many-body wave functions. J. Supercond. Nov. Magn. 33, 1697–1707 (2020)arxiv.org/abs/1912.04529
  70. [70]Oliveira, L.N., Gross, E.K.U., Kohn, W.: Density-functional theory for superconductors. Phys. Rev. Lett. 60, 2430–2433 (1988) https://doi.org/10.1103/PhysRevLett.60.2430
  71. [71]Hohenberg, P., Kohn, W.: Inhomogeneous electron gas. Phys. Rev. 136, 864–871 (1964) https://doi.org/10.1103/PhysRev.136.B864

Paper details

vlnc:W_76ph2tjc
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

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

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents