Expand this Topic clickable element to expand a topic
Skip to content
Optica Publishing Group

New method of second quantization of the strained-graphene Kerr and Faraday rotations

Open Access Open Access

Abstract

We illustrate Kerr and Faraday rotation in the strained-graphene by applying the second quantization method as an alternative approach. We consider the right- and left-going photon fields coupling with strained graphene. In other words, we have a new stationary state solution describing this phenomenon. A single-photon polarization in the provided state is considered in cases of a non-magnetic field, and uniform strained graphene. We show that the optical l properties of Faraday rotation, reflectance, and transmittance depend on the spinor phase and the energy level of an electron in strained graphene. These values can be controlled by variation of a strain parameter and strain types. Then, it is possible to have an alternative measurement of the pseudo-spin state and electronic structure in the 2-D layer by observing the optical properties of the single-photon in the provided state.

© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement

1. Introduction

Graphene, a monolayer sheet of carbon atoms, exhibits intriguing optical properties that arise from its massless Dirac dispersion of electrons [1-4]. To predict the optical properties of a material such as the optical conductivity for an out-of-plane deformation in graphene, an approach based on solutions of the Dirac equation in curved space has been used [5]. From the point of view of the novel applications of graphene, a uniform external magnetic field is applied to the system in the same direction of photon propagation, and the interaction causes a rotation of a plane of polarization after photon passing through the layer. The effect depends on the strength of the external magnetic field and the thickness of a material. The angle of rotation in a single layer of graphene is remarkable because of an enormous effect [6], which is different from other material with the same thickness. By replacing the magnetic field by a nonuniform strain field which induces a pseudomagnetic field in graphene, Faraday rotation angle π/8 for suitable terahertz waves can be achieved [7]. For the theoretical description, the famous solution to solve this problem always defines the Hamiltonian of the system in terms of interaction-free and interaction Hamiltonian. It is convenient to express Hamiltonian and eigenstates of the electron in graphene in a second quantization interacting with the classical electric field. This method is called the equation of motion (EOM) [8]. For another method, a linearized collisionless quantum kinetic equation was used to the matrix distribution function that encodes and ↓ spin-density contributions in its diagonal elements and transverse spin-density contributions in its off-diagonal elements [9]. In predictions of EOM [8], Faraday’s rotation angle

θFcμ01+ϵrRe[qσxy(q,ω)]
is obtained from the transmission amplitudes, derived by using the electromagnetic boundary conditions between two media as (D2D1)n^=ρ, (B2B1)n^=0, (E2E1)×n^=0 and n^×(D2D1)=j, ρ and j being the graphene charge and current densities, n^ unit normal at the interface. σxy is an off-diagonal component of one-photon conductivity tensor for one valley and spin. The description of the theoretical method is to consider the classical electric field, corresponding to the limit of a large number of photons. It implies that a large number of photons cause many of a completely random transition of electrons from valence band to conduction band in the material. So conductivity used in prediction of Faraday rotation contains a summation over all-state as Eq. (1). When we use this relation to calculate Faraday’s rotation angle in case of the non-magnetic field and uniform strained graphene. We will get a result of the summation in Eq. (1) to be zero. Therefore, the Faraday effect does not occur in these predictions. However, we will show that a single interband transition plays a role in understanding a single photon Faraday effect for the above cases.

For theoretical second quantization approach to the field and material interaction, many theoretical [10–13] have mainly considered the scattering properties of a single plasmon interacting with emitters as two-level quantum dots (QDs) and a V-type three-level QD. These are demonstrated that the transmission and reflection of a single plasmon can be switched on or off by controlling the detuning and changing the inter-particle distances between the QDs. The coupling between metal nanowires and quantum emitters is similar to a coupling between the polarization of light and electron in the thin-film matter.

This paper is motivated by [11]. We have investigated the quantum state of the polarization of light after passing a medium which has never been derived before. The scope of the present paper is to apply this method to describe the Faraday rotation effect in the full quantization limit by considering photon propagation in the form of the right-going and left-going photon field coupling with the uniform strained graphene and analyze in detail an alternative scheme which obtains from a new kind of stationary states solution describing Faraday rotation effect. This paper is arranged as follows. In Sec. II, we show our method that provides full second quantization describing Faraday rotation effect in the form of the right-going and left-going photon field coupling with the matter. Moreover, we propose a time-independent state of a single photon with the metal discussing an interband transition. In the next section, we consider the right and left hand circularly transmission (reflectance) for the strained graphene. The giant Faraday rotation in a single layer of graphene and uniform strained graphene can be provided. Finally, conclusions are presented in Sec. IV.

 figure: Fig. 1

Fig. 1 Show a single photon propagating in the z direction with polarization in the x direction interacts with an electron in the material. The polarization of a reflected photon and a transmitted photon is rotated in Faraday effect.

Download Full Size | PDF

2. The right-going and left-going photon-topological insulators interaction with a full second quantization approaches

In our model, we consider a single photon, propagating in the direction perpendicular to graphene. Our method, full second quantization method is discussed. Let us start with Schrödinger equation for the right and left going photon, propagating in the z direction with polarization in the η direction, (η=x,y is a photon polarization index) and interacting with an electron in material as Fig. 1

it|ψ(t)**=Hsys|ψ(t)**.
where Hsys is given as
Hsys=Hti+Hph+Hphti.

Following the Bernevig-Hughes-Zhang model [14–16], a tight-binding Hamiltonian in the q space is given

Hti=q(c^c,qc^v,q)Hti(c^c,qc^v,q),
where c and v are the band index for conduction and valence band, respectively. The single-particle Hamiltonian Hti is given as
Hti=H0(q)σ0+H1(q)σ1+H2(q)σ2,
where σ0 is the identity matrix and σ1 and σ2 are the Pauli matrices,
σ0=(1001),σ1=(0110),σ2=(0ii0),
and Hμ(q) can be expanded in the powers of the momentum q and dependent on material [14]. Solving an eigenvalue equation from Eq. (5), we obtain the eigenenergies and eigenstates as
ϵλ(q)=H0(q)+λ(H1(q))2+(H2(q))2,ψλ(q,r)=12(1λeiλϕq)eikr,
where ϕq=tan1[H2(q)/H1(q)] is the phase of spinor and λ=±1 is the band index for conduction and valence band, respectively.

For the uniform strain formulation, the effects of deforming the lattice are taken into account, leading to a modification of the hopping. Modification of the hopping energies between different sites will, in turn, lead to new terms in the original Hamiltonian in the tight-binding formulation. Here the sum is over all nearest-neighbor pairs t(|δij|) having the strained hopping energy due to strain. The length and direction of the three nearest-neighbor vector δi transform under strain according to δi(1+ε)δ0i where δ0i represents a nearest-neighbor vector in the non-deformed graphene plane, and ε is the strain tensor [17],

ε=(εxxεxyεxyεyy).

For Cartesian strain tensor, we assume the x axis is along graphene armchair direction and the y axis is along graphene’s zigzag, respectively. In reciprocal space, the variation of hopping amplitude with inter carbon distance is t(|δij|)=t0exp [β(|δij|/a1)] with a=1.42Ao being the unstained nearest-neighbor separation, t0=2.7eV, and β3. Then the hopping perturbation δti can be obtained by expanding t(|δij|) and keeping the first order,

t(a+Δδi)t0(1(β/a)(|(1+ε)δ0iδ0i|)),
with Δδi as the variation of bond length. Most importantly, the symmetry point K=4π33(a,0) and K=4π33(a,0) that coincides with the Fermi point in the unstrained, equilibrium situation, and chosen here for definiteness moves to the new Kδk, (δk the shift in the Dirac point)
δk=πa(433εyy+β2π(εxxεyy),433εxy+βπεxy),
and K is shifted in the opposite direction, δk. The shape of the Fermi surface contour changed due to the strain types are shown in [17]. As can be seen, the figures show that the lattice deformations in the calculation of the vector δk generate a shift of the Dirac points in reciprocal space to accurately describing the strain-dependent. Introducing Hμ(q) for uniform strain are given [17–19] as
H0(q)=0,,H1(q)=v1qvFecApdx ,H2(q)=v2qvFecApdy
when c is the speed of light and the pseudovector potential is given as Apd=(Apdx,Apdy)=τceδk with τ=±1 corresponding to K and K, respectively. The velocities vμ=(vμx,vμy) are obtained in terms of the matrix elements of strain tensor, εxx, εxy and εyy as
v1y=vF(114β(εxx+3εyy)2π33εyy),v1x=τvF(12βεxy2π33εxy) 
and
v2y=vF(12β(εxy2π33εxy),v2x=τvF(114β(3εxx+εyy)+2π33εyy).

When the strain tensor vanish, the two velocities reduce to v1=(0,vF), v2=(τvF,0) and Apd=0, Eq. (11) then becomes the Dirac Hamiltonian.

For the non-uniform strained graphene [7], the lattice distortion is (ux,uy)=(xy/R,x2/2R); R is the bending radius of the displacement applied, and the strain tensor is related to the displacement fields by εxx=y/R, εyy=0 and εxy=0. The presence of a strain-induced vertical magnetic field (Bpd=cβπ/aR(0,0,1)) cause the cyclotron motion of the electron. A single interband transition plays a role in setting the Hall conductivity, even for zero Fermi energy and the Hall conductivity of graphene σxy can be obtained as Eq. (1) of [7]. In our work, we are interested in the uniform strain (εxx, εyy and εxy are constants) then the uniform pseudo vector potential appears in the material and produces a zero pseudo magnetic field (Bpd=0). The uniform strain only changes Fermi line from the isotropic circle to an elliptical shape due to the applying strains [18].

Following the method proposed in [10–13], Hamiltonian of the quantized electromagnetic field, moving along the z direction is rewritten as

Hph=icη=x,ydz[bL(η)(z)zbL(η)(z)bR(η)(z)zbR(η)(z)]
when operations bR(η)(z) and bL(η)(z) are respectively bosonic operators creating a right- and left-going photon at z with polarization in the η direction,
bR(η)(z)=0dkeikzbk(η)andbL(η)(z)=0dkeikzbk(η),
where bk(η) and bk(η) are the bosonic annihilation and creation operator of the propagating photon mode k with the frequency of the photon ωk in the direction η. The coupling between the propagating photons and the material is given as
Hphti=γ=R,Lη=x,ydzνηδ(z)[bγ(η)(z)Jη+bγ(η)(z)Jη]
where νη=2π/ϵrωkA0 is the coupling constant and A0 is the unit cell area. The quantity ϵr is the relative electric constant. The current operator Jη in the η direction is given as
Jη=qλ=c,vλ=c,vψλ|jη|ψλc^λ,qc^λ,q.
where jη=eHti/qη is the relativistic current density.

To investigate Schrödinger for the right and left going photon, we introduce a time-independent state of the system consisting of linearly polarized photon state and a transition state in the matter

|ψ(q,ϵk)=γ=R,Lη=x,y dzEk,γ(η)(z)bγη(z)|0,μ+ek(η)(ϵq)P^c,v(η)(q)|0,μ,
where |0,μ is the vacuum state of photon and a quantum state of an electron in the matter with a state index μ. The generic operator or the dipole operator is given as
P^c,v(η)(q)=ψc|jη|ψvc^c,qc^v,q,
corresponding to a one-by-one transition from an initial state in valence band to a final state in conduction band by coupling with a linearly polarized photon in the η direction. In transitions of the electrons, the energy and the momentum and the spin are conserved, respectively. These transitions connect states in different bands, they are referred to as being inter-band-like. ek(η)(ϵq) is the probability amplitude of the electron in the final state. The quantum state Eq. (reftime independent state) implies consistent histories, which is a way to expresses a quantum history states such as [12] and [20]. For a single-photon state, we can write the polarization state of a photon as a superposition of any basis polarization states as linearly or ellipsoidally polarized states. In Eq. (18), the photon state is in the linear polarization basis states where Ek,R(L)(η)(z) is respectively an amplitude of finding the right “R” and left “L” going photon having the η linearly polarization at the z position.

Assuming Ek,±(η)=Ek,R(η)±Ek,L(η) the linearly plus and minus fields with the η direction and applying the time independent Schrödinger equation Hsys|ψ(q,ϵk)=nk|ψ(q,ϵk), we obtain the equations of motion :

(ϵkϵμ)Ek,+(η)(z)+iczEk,(η)(z)νηδ(z)[ek(η)(q)η,η+ek(η)(q)η,η]=0
and
η=x,y(ϵkη,ηΛη,η)ek(η)(q)νη(Ek,+(η)(0)η,ηEk,+(η)(0)η,η)=0
when η,η=μ|JηP^c,v(η)(q)|μ and Λη,η=μ|JηHtiP^c,vη(q)|μ. For calculating η,η and Λη,η, we assume that Hti|μ=ϵμ|μ is an energy eigenvalue of the electron in material state μ at zero temperature. We can calculate the generic- current correlation η,η for keeping only the first order approximation of occupation number in conduction nc(v)(q) and taking nc(q)=0 and nv(q)=1 for zero temperature
η,η=ψv|jη|ψcψc|jη|ψvnv(q)
and Λη,η=μ|JηHtiP^c,vη(q)|μ is obtained as
Λη,η=ψv|jη|ψcψc|jη|ψv(ϵμ+ϵcϵv)nv(q).

Then the combination of relationship in Eqs. (21)-(22) can be expressed in the form of Maxwell’s equation of electric field

2z2Ek,+(η)(z)+iωμ0δ(z)η=x,yση,ηEk,+(η)(z)+ω2εrμ0Ek,+(η)(z)=0;ωk=ϵkϵμ,
which Ek,+(η)(z) is the electromagnetic field with linearly polarized along the η direction and the minimal optical conductivity tensor ση,η is rewritten as
ση,η=(i4πϵrc2A0ωk)[x,xx,y  y,xy,y][ϵkx,xΛx,xϵkx,yΛx,y  ϵky,xΛy,xϵky,yΛy,y]1[x,xx,y  y,xy,y].

By considering the one to one interaction and having an electron transition from the conduction band to a final state in the valence band, one-photon conductivity tensor for one valley and spin ση,η is obtained in the form

ση,η=(πϵrc2A0)ψv|jη|ψcψc|jη|ψviωk(ωk(ϵcϵv))nv(q).

Eq. (26) shows the resonance preserved in the optical conductivity tensor as a function of each value of photon energy ωk which is a resonance energy eigenvalues. This optical conductivity corresponds with the optical conductivity investigated by an effective quantum field theory graphene sheet with arbitrary one-dimensional strain field from a microscopic effective low energy Hamiltonian at zero temperature [21]. The distortion of the reciprocal lattice transforms the original Dirac cone into a distorted one with a directional-dependent Fermi velocity. The Fermi line deforms from the isotropic circle into a different ellipse between two valleys, K and K. This is due to the fact that in the presence of the strain, we have ϵK,λ(q)ϵK,λ(q). So the conductivity tensor in Eq. (26) no satisfy the relation ση,η(K)(q)=ση,η(K)(q) for inequivalent valleys. Faraday’s rotations demonstrated by our approach will not cancel each other after we consider both valleys.

From Maxwell’s equation for electric field Eq. (24) and the boundary conditions at graphene with the air interface, the continuity of the tangential components of the electric field at the surface of graphene (z=0) is,

(Ek,+(x,i)(0),Ek,+(y,i)(0))+(Ek,+(x,r)(0),Ek,+(y,r)(0)))=(Ek,+(x,t)(0),Ek,+(y,t)(0))
and the derivatives are evaluated at z = 0
zEk,+(η,t)(0)zEk,+(η,i)(0)zEk,+(η,r)(0)=iωμ0η=x,yση,ηEk,+(η)(0)
where the last condition was derived from integrating Eq. (28) in the interval z[0,0+]. Using algebra and circularly polarized coordinates, Eqs. (27) - (28), we therefore obtain the transmitted and reflected fields in the form of the minimal optical conductivity tensor, Γμν=cμ0σμν/2 as
[Ek,+(x,t)(0) Ek,+(y,t)(0)]=[(1+Γxx)Γxy  Γyx(1+Γyy)]1[Ek,+(x,i)(0) Ek,+(y,i)(0)],
and
[Ek,+(x,r)(0) Ek,+(y,r)(0)]=[(1+Γxx)Γxy  Γyx(1+Γyy)]1[ΓxxΓxy  ΓyxΓyy][Ek,+(x,i)(0) Ek,+(y,i)(0)].

To simplify the problem without loss of the generality, one can assume the polarization state of incoming photon by setting Ek,+(y,i)=0. The right (left) hand circularly transmission t+(t) and reflections r+(r) coefficients are respectively given as

r±=|r±|eiθK(±)=|(Ek,+(x,r)±iEk,+(y,r))/Ek,+(x,i)|=±iGΠx(Πx±iΠy)2ωk(ωk(ϵcϵv)+iG(|Πx|2+|Πy|2)
and
t±=|t±|eiθF(±)=|(Ek,+(x,t)±iEk,+(y,t))/Ek,+(x,i)|=ωk(ωk(ϵcϵv)±GΠy(Πx±iΠy)2ωk(ωk(ϵcϵv)+iG(|ix|2+|Πy|2)
where G=e2π/2ω0A0 and ω0=cϵr/μ0 and Πη,η=x,y is given as
Πη=±i(v1η)2+(v2η)2sin(ϕqϕq,η(s)),ϕq,η(s,τ)=tan1(v2η/v1η).

Optical transition matrix elements Eq. (33) contain the effects of the minimal-coupling between the light and matter that describe the possible transitions allowed by the spin selection rules [22]. The conservation of the electron momentum is in a plan of the strained graphene, and the pseudo-spin flip creates a change of the spin angular momentum of light when the transition appears between the positive and negative energy bands. Since an initial and final momentum point in the same direction and have the same magnitude; i.e., the transitions are vertical. According to Eq. (33) the maximum coupling of the photon and the spinor have to along the direction satisfying condition ϕqϕq,η(s,τ)=±π/2. Phase shifts ϕq,η(s,τ) determined by Eq. (33) give the same value for K and K valleys corresponding with velocities in Eqs. (12)-(13). We will discuss this result in the next section.

We want to point out the connection between polarization ratio ty/tx and the general velocity matrix elements between the conduction band and the valence band Πy and Πx as

|tytx|=|2(ωk(ωkϵcv)(cμ0/A0)ΠyΠy(cμ0/A0)ΠyΠx|,ϵcv=ϵcϵv.

For near resonance frequency ωkϵcv, the polarization ratio come to |ty/tx||Πy/Πx|, corresponding with the experimentally observed PL polarization ratio and the quantum mechanical pseudospin degrees of freedom in graphene [23]. Transmittance T(ω) and reflectance R(ω) can be written as

T(ω)=12(|t+|2+|t|2),  R(ω)=12(|r+|2+|r|2)
where the factor 1/2 comes from the proper normalization of circularly polarized waves. Faraday angle θF and Kerr angle θK are respectively given by
θF(ω)=12(θF(+)θF()),  θK(ω)=12(θK(+)θK()).

Let us consider a different between a classical electromagnetic field and a second quantization approach. We start with the scattered state, which can be expressed as |out=nS(n)|in, where S(n) is the S-matrix operator of n incident photons, whose elements can be calculated using scattering theory. For example, the one-mode one-photon S-matrix S(1) can be written as

S(1)=ηdktηbR,kη|0,μ0,μ|bR,kη+dkrηbL,kη|0,μ0,μ|bL,kη.

In experiments, the statistics of scattered photons is predominately determined by measuring first-order correlation gη,η(1)(z1,z2) of the scattered fields. For the transmitted beam, it is defined as

gη,η(1)(z1,z2)=out|bRη(z1)bRη(z2)|outout|bRη(z)bRη(z)|out.

For a classical electromagnetic field approach, the first-order correlation is given as

gη,η(1)cl(z1,z2)=dkEk,+(η)(z2)Ek,+(η)*(z1)dkEk,+(η)(z)Ek,+(η)*(z)=dktη(k)tη*(k)eik(z1z2)dktη(k)tη*(k).

For a quantum phenomenon, we start with a non-coupling of a single photon initial state |in then we can obtain the scattered state as |out=S|in, which contains many-body effects of the photon in the system. Two possible outcomes exist, Eq. (38) can be expanded as

gη,η(1)(z1,z2)=in|S(1)bRη(z1)bRη(z2)S(1)|in+n>1in|S(n)bRη(z1)bRη(z2)S(n)|in.

The first term of Eq. (40) corresponds with the classical electromagnetic field first-order correlation Eq. (39). It implies that a single photon is due to a transition of an electron from the initial state in the conduction band to a possible final state in the valence band. Thus we can obtain a classical electromagnetic field approach in Faraday effect from a very large number of the photon to be in the time-independent state Eq. (18) and summing a completely random transition of electrons in the material. Next, the second term contains many photon effects, which transport through strained graphene. This effects had been considered as photon-photon correlations and entanglement generation in a one-dimensional waveguide coupled to two qubits with an arbitrary spatial separation [11, 12].

3. Calculated results

In this section, we consider the giant Faraday rotation in graphene as discussed by the exact solution of the time-independent Schrödinger equation for the light-material interaction Eq. (18). Two cases such as non-magnetic field and uniform strained graphene is moderated, deformation <20%. This allows the hopping energy around its non-deformed value t, and t/a5eV. The reciprocal space is shifted from the traditional Dirac points. The nature of the contours of the strained band structure has been discussed in [17]. For non-uniform strain, the Landau level structure is described as one modified by a uniform effective magnetic field [21] and found that the field is strongest when the force is applied in the direction perpendicular to armchair orientation. For strongly deformed lattice [24], the effect of the hopping mechanism for the next-nearest neighbors leads to anti-symmetric properties of the energy spectrum around zero energy. Furthermore, the robust deformation causes unusual dispersion relation which is linear in one direction and quadratic in the other [25, 26].

3.1. Kerr rotation and Faraday rotation-dependent on phase’s spinor and energy level

For example cases, we consider linearly-polarized light interacts with graphene for the non-magnetic field, and uniform strained graphene, the polarization state of the light will change in two ways. The first way involves the rotation of the polarization axis, and the second is a change from linearly-polarized to elliptically-polarized light that is characterized by the ellipticity. The phase (amplitude) variations between the right and left-circularly-polarized light components is responsible for the rotation, denoted by transmission t+(t) and reflectance r+(r) coefficients. We show that these characteristics correspond to the phase of spinor and energy level. The phenomenological and analytical expressions in such cases can be found in the next studies.

 figure: Fig. 2

Fig. 2 Reflectance R and Kerr rotation θK for single electron interacting with single photon by varying q ≡ (qx, qy) in non-strained graphene. A red line for electron with wave vector (1,7.53) ×108 m−1, blackline for (5,5.72) ×108 m−1, blue line for (1,7.53) ×107 m−1 and green line for (5,5.72) ×107 m−1 around the K valley (τ = 1). Insets in Fig. 2(b) show the plots of imaginary part which are corresponding to the critical conditions.

Download Full Size | PDF

Reflectance and Kerr rotation : Refer to Eqs. (31)-(32), real parts and imaginary parts of left handed and right handed circular polarization are the keys to explain the optical properties, i.e. reflectance and Kerr rotation. Let usconsider four different wave vectors around the K valley for graphene such as, q1=(1,7.53)×108 m 1, q2=(5,5.72)×108 m 1, q3=(1,7.52)×107 m 1 and q4=(5,5.72)×107 m 1 and the energies for first two wave vector is 0.5 eV and another two wave vectors are 0.05 eV, respectively. The corresponding phases of electronic spinors ϕq for each energy level are 1.44 and 0.85 radian.

The phase of spinor ϕq controls the characteristic curve of reflectance and photon energy ωk. In Fig. 2(a), there are two resonant peaks of reflectance. The first one is from ω to 0 and another one point, photon energy is equal to ϵcv. Focusing on the low energy level in Fig. 2(a), the reflectance seems to be constant (blue and green lines) because of the merging of two resonant peaks. In higher energy level, the peaks split apparently as in the red and black lines. However, it appears that the magnitude of reflectance is independent of the energy level. For examples, As increasing the phase of pseudo-spin state ϕq near π/2 for high energy level and low energy level, the reflectance R reaches the maximum, according to the relation of Eq. (33). The magnitude of the curve is determined primarily by the phase of the pseudo-spin state.

In Fig. 2(b), the plots of Kerr rotation depends on the phase of spinor ϕq and energy level ϵc(q). The behaviors are also similar to reflectance. As increasing photon energy, there is the merging of 2 barriers for low energy level and ϕq near π/2 (green line). The peaks split apparently as in the red and black lines for high energy level. Another characteristic, Kerr rotation angle changes rapidly by π when photon frequency has a value around the specific frequencies. At the specific frequencies, the phase of the right (left) handed circularly reflected-photon is the same as that phase of the incident photon. The frequencies that change the polarization of reflected photon are called the critical frequencies ωc(±). The critical frequencies can be found by the conditions Im[r+]=0 and Im[r]=0 or argument of right-handed (r+) and left-handed (r) circular polarization to be zero (θK(±)=0) [see Eq. (31)] as

ωc(±)=12(ϵcv±(ϵcv)24  G (Πx/Πy)( |Πy|2+|Πx|2)).

These are valid only for ω>0. From Eq. (31) and Eq. (33), the responses peaks of reflectance is invariant under the parity transformation, changing sign (positive or negative) and keeping the same magnitude of the wavevector, i.e. (qx,qy)(qx,qy). Because the r+ and r are interchangeable, the reflectance and the Kerr rotation is still unchanged (symmetric) under the parity transformation.

 figure: Fig. 3

Fig. 3 Transmittance T and Faraday rotation θF for single electron interacting with single photon by varying q ≡ (qx, qy) in non-strained graphene. A red line for electron with wave vector (1,7.53) ×108 m−1, black line (5,5.72) ×108 m−1, blue line (1,7.53) ×107 m−1 and green line (5,5.72) ×107 m−1 around the K valley (τ = 1). Insets in Fig. 3(b) show the plots of imaginary part which are corresponding to the critical conditions.

Download Full Size | PDF

Transmittance and Faraday rotation : The behavior of the transmitted photon corresponds to the reflected photon. It quite completely transparent except for the near-resonant region which implies the minimal coupling between the light and material. The lowest transmittances correspond with ωk to 0 and ϵcv because of the maximum reflectance. When the phase of spinor decreases from π/2, by comparison, the transmittances are not too low as shown in Fig. 3(a) which the black lines are above the red lines (also for green and blue lines). The minimum transmittances are close to each other and merge for low energy level consistent with the reflectance. As an electronic energy level increasing, the transmittance split apparently (the red and black lines). This characteristic is seen in a curve of Faraday rotation too as shown in Fig. 3(b). Transmittance and Faraday rotation is invariant under the parity transformation which is agreeable with reflectance and Kerr rotation. The reasons are that deformation as shear strain, uniaxial armchair strain, and zigzag strain give the ellipse Dirac cone changing in the opposite direction of reciprocal lattice parameters against the real lattice parameters.However, the ellipse Dirac cone is invariant under the parity transformation.

3.2. Kerr rotation and Faraday rotation-dependent on strain types

Phase’s spinnor: We consider strain tensors as εis tensile isotropic strain, εss shear strain, εas armchair uniaxial strain, and εzs zigzag strain, respectively. Here ν is Poisson’s ratio and ε is a strength of the strain. Each strain tensors is given as

εis=(ε00ε), εss=(0εε0), εas=(ε00νε), εzs=(νε00ε).

 figure: Fig. 4

Fig. 4 Show that transmittance T and reflectance R as a function of the phase of spinor for q = (5,3)×108 m−1 around the K valley (τ = 1) and Poisson’s ratio ν = 0.3 by varying strain parameter in different strain types compared with non strain case (black line). Inset in Fig. 4(b) show that the phase shift ϕq,η(s,τ),η=x,y for shear strain type as a function of stain parameter.

Download Full Size | PDF

We calculate transmittance T and reflectance R as a function of the phase’s spinor for each strain type comparing with a non-strain case, evidently in Fig. 4(a). These results show that transmittances increase from non-strained-graphene (ε = 0, black line) for tensile isotropic strain (blue line) and zigzag strain (red line) with a strength of the strain parameter ε=0.1. In the opposite case, the transmittance for armchair uniaxial strain (green line) decreases from the black line. For shear strain, the orange line is almost overlapping with the black line. As expected for reflectance, this characteristic curve is in the opposite of transmittance curve, (T+R=1), see in Fig. 4(b). We consider the effects of strain on the phase shift’s spinor. From Eqs. (12)-(13) for tensile isotropic strain, zigzag strain, and armchair uniaxial strain, the phase shift’s spinor ϕq,y(x)(s,τ) is 0 (π/2) for all strain parameter, respectively. For shear strain case, the phase shift’s spinor has a slight change from the value 0 to π/40 and in opposite from the value π/2 to π/2+π/20 when ε=0.2, for ϕq,y(s,τ) and ϕq,x(s,τ) respectively, showed inset in Fig. 4(b). So the maximum and minimum point of transmittance and reflectance for each strain type are near the phase’s spinor 0, 2π (maximum) and π/2 (minimum). These results give the same values for K and K.

 figure: Fig. 5

Fig. 5 Reflectance R and Kerr rotation θK (in radian) for q = (5,3)×108 m−1 around the K valley (τ = 1) by varying strain parameter in different strain types. Using Poisson’s ratio ν = 0.3 and ε = 0 (red), ε = 0.05 (black) and ε = 0.1 (blue). Inset in Figs. 5(a), 5(c), show a little shift to the left corresponding to the energy level changing with strain.

Download Full Size | PDF

 figure: Fig. 6

Fig. 6 Transmittance T and Faraday rotation θF (in radian) for q = (5,3)×108 m−1 around the K valley (τ = 1) by varying strain parameter in different strain types. sing Poisson’s ratio ν = 0.3 and ε = 0 (red), ε = 0.05 (black) and ε = 0.1 (blue). Inset in Figs. 6(a) and 6(c), show a little shift to the left corresponding to the energy level changing with strain.

Download Full Size | PDF

Reflectance and transmittance in strained-graphene: As applying force to graphene layer, the effects of strain on the change of the bond lengths are displayed by reflectance, transmittance, Kerr and Faraday rotation. We show the results of the optical response of 4 strain types as armchair, zigzag, shear, and isotropic strain. We use Poisson’s ration ν=0.3 for armchair and zigzag strain types. The selected-wave vector is q=(5,3)×108 m 1 and selected-strain parameters are ε = 0, ε=0.05, and ε=0.1. Evidently in Figs. 5-6, the peaks shift to the left for all of the strain types. They split clearly for isotropic (Figs. 5(b) and 5(d)) and zigzag types (Figs. 5(d) and 5(h)) while the splitting is much less for shear (Figs. 5(a) and 5(c)) and armchair strain types (Figs. 5(c) and 5(g)). The maximum angles are the same but a shift to the left corresponding to the energy level changing with strain. Evidently, the corresponding energies ϵcv of zigzag strain are 0.384, 0.346, and 0.313 eV and of armchair strain are 0.384, 0.376, and 0.371 eV, as seen in inset of Figs. 5(a) and 5(c) for reflectance and Figs. 6(a) and 6(c) for transmittance cases of each strain parameters, respectively. The reflectances and transmittances in Figs. 5 - 6 show the similar results as that in [13] dealing with transition energy. For phases of the spinor, strain parameters have a little change on the phases of the spinor such as zigzag strain type as 0.540, 0.648, and 0.780 radians, and in armchair strain types as 0.540, 0.480, 0.417 radians for each parameter ε, respectively.

4. Discussion and conclusions

In this paper, we showed the alternative way to consider Kerr and Faraday rotation for two-dimensional materials, which is a wave function-like description of a Dirac fermion with a second quantization approaches and presented applications of the theory and technique. We treat the incoming photon field in the second quantized form, allowing this way to obtain the classical result [8] in the limit of a large number of photons, due to a completely random transition of electrons in graphene. For the summation of the overall transition of electrons, Eqs. (22) - (23) come to a diagonal matrix and get Faraday angle to be zero. In this result, it corresponds to the relation in Eq. (1) to calculate Faraday’s rotation angle. Therefore, the Faraday effect does not occur in these predictions. However, we show that a single interband transition plays a role in understanding a single photon Faraday effect. For examples, we demonstrate the method for the cases of isotropic, shear, armchair uniaxial and zigzag strains. The Faraday rotations in the infrared regime are generated and measurable Faraday rotation angles. Reflection and transmission coefficients dependent on strain parameters and types are shown. An extremum of reflectance and transmittance exist at near the phase of spinor π/2, and the resonance peaks are found at ωkϵcv. The Kerr rotation switches suddenly by π when the imaginary part of the reflection coefficient equals to zero (critical frequencies ωc(±)). These show that the optical properties can be tuned by using strain control. To manipulate the geometrical phase of the electron by using strain control, a single photon Faraday effect can be operated as a switch to control the transmittance of the photon inside the quantum network and as a variable phase of a single photon qubit or the photon’s polarization rotation by varying strain parameter. In another way, it can be investigated by observing the Kerr and Faraday rotation process to study the band structure and the phase of the spinor of two-dimensional materials.

Funding

Kasetsart university research and development institute (KURDI), Specialized center of rubber and polymer materials in agriculture and industry (RPM), and Faculty of Science, Kasetsart University.

Acknowledgments

We acknowledge helpful discussions with I. M. Tang and Papichaya Chaisakul.

References

1. A. H. Castro Neto, F. Guinea, N. M. R. Peres, K. S. Novoselov, and A. K. Geim, “The electronic properties of graphene,” Rev. Mod. Phys. 81, 109 (2009). [CrossRef]  

2. N. M. R. Peres, “The transport properties of graphene: An introduction,” Rev. Mod. Phys. 82, 2673 (2010). [CrossRef]  

3. R. Shimano, G. Yumoto, J. Y. Yoo, R. Matsunaga, S. Tanabe, H. Hibino, T. Morimoto, and H. Aoki, “Quantum Faraday and Kerr rotations in graphene,” Nat. Commun. 4, 1841 (2013). [CrossRef]   [PubMed]  

4. K. N. Okada, Y. Takahashi, M. Mogi, R. Yoshimi, A. Tsukazaki, K. S. Takahashi, N. Ogawa, M. Kawasaki, and Y. Tokura, “Terahertz spectroscopy on Faraday and Kerr rotations in a quantum anomalous Hall state,” Nat. Commun. 7, 12245 (2016). [CrossRef]   [PubMed]  

5. A. J. Chaves, T. Frederico, O. Oliveira, W. de Paula, and M. C. Santos, “Optical conductivity of curved graphene,” J. Phys.: Condens. Matter 26, 185301 (2014).

6. I. Crassee, J. Levallois, A. L. Walter, M. Ostler, A. Bostwick, E. Rotenberg, T. Seyller, D. van der Marel, and A. B. Kuzmenko, “Giant Faraday rotation in single- and multilayer graphene,” Nat. Phys. 7, 48 (2010). [CrossRef]  

7. J. C. Martinez, M. B. A. Jalil, and S. G. Tan, “Giant Faraday and Kerr rotation with strained graphene,” Opt. Lett. 37, 3237 (2012) [CrossRef]   [PubMed]  

8. A. Ferreira, J. Viana-Gomes, Y. V. Bludov, V. Pereira, N. M. R. Peres, and A. H. C. Neto, “Faraday effect in graphene enclosed in an optical cavity and the equation of motion method for the study of magneto-optical transport insolids,” Phys. Rev. B 84, 235410 (2011). [CrossRef]  

9. W-K. Tse and A. H. MacDonald, “Giant magneto-optical Kerr effect and universal Faraday effect in thin-film topological insulators,” Phys. Rev. lett. 105, 057401 (2010). [CrossRef]   [PubMed]  

10. D. E. Chang, A. S. Sorensen, A. E. Demler, and M. D. Lukin, “A single-photon transistor using nanoscale surface plasmons,” Nat. Phys. 3, 807 (2007). [CrossRef]  

11. M. Laakso and M. Pletyukhov, “Scattering of two photons from two distant qubits: exact solution,” Phys. Rev. lett. 113, 183601 (2014). [CrossRef]   [PubMed]  

12. P. Srisangyingcharoen, R. Klinkla, and S. Boonchui, “Dynamics of propagating surface plasmon induced photon emission from quantum dots: quantum history approach,” J. Phys. B: At. Mol. Opt. Phys. 48, 215501 (2015). [CrossRef]  

13. J. T. Shen and S. Fan, “Coherent photon transport from spontaneous emission in one-dimensional waveguides,” Opt. Lett. 30, 2001 (2005). [CrossRef]   [PubMed]  

14. C-X. Liu, X-L. Qi, H. Zhang, X. Dai, Z. Fang, and S-C. Zhang, “Model hamiltonian for topological insulators,” Phys. Rev. B 82, 045122 (2010). [CrossRef]  

15. B. A. Bernevig, T. L. Hughes, and S. C. Zhang, “Quantum spin hall effect and topological phase transition in HgTe quantum wells,” Science 314, 1757 (2006). [CrossRef]   [PubMed]  

16. Z. B. Siu, S. G. Tan, and M. B. A. Jalil, “Effective Hamiltonian for surface states of topological insulator nanotubes,” Sci. Rep. 7, 45350 (2017). [CrossRef]   [PubMed]  

17. A. L. Kitt, V. M. Pereira, A. K. Swan, and B. B. Goldberg, “Lattice-corrected strain-induced vector potentials in graphene,” Phys. Rev. B 85, 159909 (2012). [CrossRef]  

18. P. Fanbanrai, A. Hutema, and S. Boonchui, “Effects of strain on the Schwinger pair creation in graphene,” Physica B 472, 84 (2015). [CrossRef]  

19. V. M. Pereira and A. H. C. Neto, “Tight-binding approach to uniaxial strain in graphene,” Phys. Rev. B 80, 045401 (2009). [CrossRef]  

20. D. Georgiev and E. Cohen, “Probing finite coarse-grained virtual Feynman histories with sequential weak values,” Phy. Rev. A 97, 052102 (2018). [CrossRef]  

21. S-J. Zhang, H. Pan, and H-L. Wang, “Frequency dependent optical conductivity of strained graphene at T=0 from an effective quantum field theory,” Physica B 511, 80 (2017). [CrossRef]  

22. H. P. Paudel and M. N. Leuenberger, “Giant Faraday effect due to Pauli exclusion principle in 3D topological insulators,” J. Phys.: Condens. Matter 26, 082201 (2014).

23. T. Danz, A. Neff, H. J. Gaida, R. Bormann, C. Ropers, and S. Schäfer, “Ultrafast sublattice pseudospin relaxation in graphene probed by polarization-resolved photoluminescence,” Phy. Rev. B 95, 241412 (2017). [CrossRef]  

24. R. O. Dillon, I. L. Spain, and J. W. McClure, “Electronic energy band parameters of graphite and their dependence on pressure, temperature and acceptor concentration,” J. Phys. Chem. Solids. 38, 635 (1997). [CrossRef]  

25. P. Dietl, F. Piechon, and G. Montambaux, “New magnetic field dependence of landau levels in a graphenelike structure,” Phys. Rev. lett. 100, 235405 (2008). [CrossRef]  

26. T. Thitapura, W. Liewrian, T. Jutarosaga, and S. Boonchui, “Curvature effect on polarization of light emitted from chiral carbon nanotubes,” Opt. Express 25, 25588 (2017). [CrossRef]   [PubMed]  

Cited By

Optica participates in Crossref's Cited-By Linking service. Citing articles from Optica Publishing Group journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (6)

Fig. 1
Fig. 1 Show a single photon propagating in the z direction with polarization in the x direction interacts with an electron in the material. The polarization of a reflected photon and a transmitted photon is rotated in Faraday effect.
Fig. 2
Fig. 2 Reflectance R and Kerr rotation θK for single electron interacting with single photon by varying q ≡ (qx, qy) in non-strained graphene. A red line for electron with wave vector (1,7.53) ×108 m−1, blackline for (5,5.72) ×108 m−1, blue line for (1,7.53) ×107 m−1 and green line for (5,5.72) ×107 m−1 around the K valley (τ = 1). Insets in Fig. 2(b) show the plots of imaginary part which are corresponding to the critical conditions.
Fig. 3
Fig. 3 Transmittance T and Faraday rotation θF for single electron interacting with single photon by varying q ≡ (qx, qy) in non-strained graphene. A red line for electron with wave vector (1,7.53) ×108 m−1, black line (5,5.72) ×108 m−1, blue line (1,7.53) ×107 m−1 and green line (5,5.72) ×107 m−1 around the K valley (τ = 1). Insets in Fig. 3(b) show the plots of imaginary part which are corresponding to the critical conditions.
Fig. 4
Fig. 4 Show that transmittance T and reflectance R as a function of the phase of spinor for q = (5,3)×108 m−1 around the K valley (τ = 1) and Poisson’s ratio ν = 0.3 by varying strain parameter in different strain types compared with non strain case (black line). Inset in Fig. 4(b) show that the phase shift ϕ q , η ( s , τ ) , η = x , y for shear strain type as a function of stain parameter.
Fig. 5
Fig. 5 Reflectance R and Kerr rotation θK (in radian) for q = (5,3)×108 m−1 around the K valley (τ = 1) by varying strain parameter in different strain types. Using Poisson’s ratio ν = 0.3 and ε = 0 (red), ε = 0.05 (black) and ε = 0.1 (blue). Inset in Figs. 5(a), 5(c), show a little shift to the left corresponding to the energy level changing with strain.
Fig. 6
Fig. 6 Transmittance T and Faraday rotation θF (in radian) for q = (5,3)×108 m−1 around the K valley (τ = 1) by varying strain parameter in different strain types. sing Poisson’s ratio ν = 0.3 and ε = 0 (red), ε = 0.05 (black) and ε = 0.1 (blue). Inset in Figs. 6(a) and 6(c), show a little shift to the left corresponding to the energy level changing with strain.

Equations (42)

Equations on this page are rendered with MathJax. Learn more.

θ F c μ 0 1 + ϵ r R e [ q σ x y ( q , ω ) ]
i t | ψ ( t ) ** = H s y s | ψ ( t ) ** .
H sys = H ti + H ph + H ph ti .
H ti = q ( c ^ c , q c ^ v , q ) H ti ( c ^ c , q c ^ v , q ) ,
H ti = H 0 ( q ) σ 0 + H 1 ( q ) σ 1 + H 2 ( q ) σ 2 ,
σ 0 = ( 1 0 0 1 ) , σ 1 = ( 0 1 1 0 ) , σ 2 = ( 0 i i 0 ) ,
ϵ λ ( q ) = H 0 ( q ) + λ ( H 1 ( q ) ) 2 + ( H 2 ( q ) ) 2 , ψ λ ( q , r ) = 1 2 ( 1 λ e i λ ϕ q ) e i k r ,
ε = ( ε x x ε x y ε x y ε y y ) .
t ( a + Δ δ i ) t 0 ( 1 ( β / a ) ( | ( 1 + ε ) δ 0 i δ 0 i | ) ) ,
δ k = π a ( 4 3 3 ε y y + β 2 π ( ε x x ε y y ) , 4 3 3 ε x y + β π ε x y ) ,
H 0 ( q ) = 0 , , H 1 ( q ) = v 1 q v F e c A p d x   , H 2 ( q ) = v 2 q v F e c A p d y
v 1 y = v F ( 1 1 4 β ( ε x x + 3 ε y y ) 2 π 3 3 ε y y ) , v 1 x = τ v F ( 1 2 β ε x y 2 π 3 3 ε x y )  
v 2 y = v F ( 1 2 β ( ε x y 2 π 3 3 ε x y ) , v 2 x = τ v F ( 1 1 4 β ( 3 ε x x + ε y y ) + 2 π 3 3 ε y y ) .
H ph = i c η = x , y d z [ b L ( η ) ( z ) z b L ( η ) ( z ) b R ( η ) ( z ) z b R ( η ) ( z ) ]
b R ( η ) ( z ) = 0 d k e i k z b k ( η ) a n d b L ( η ) ( z ) = 0 d k e i k z b k ( η ) ,
H ph ti = γ = R , L η = x , y d z ν η δ ( z ) [ b γ ( η ) ( z ) J η + b γ ( η ) ( z ) J η ]
J η = q λ = c , v λ = c , v ψ λ | j η | ψ λ c ^ λ , q c ^ λ , q .
| ψ ( q , ϵ k ) = γ = R , L η = x , y   d z E k , γ ( η ) ( z ) b γ η ( z ) | 0 , μ + e k ( η ) ( ϵ q ) P ^ c , v ( η ) ( q ) | 0 , μ ,
P ^ c , v ( η ) ( q ) = ψ c | j η | ψ v c ^ c , q c ^ v , q ,
( ϵ k ϵ μ ) E k , + ( η ) ( z ) + i c z E k , ( η ) ( z ) ν η δ ( z ) [ e k ( η ) ( q ) η , η + e k ( η ) ( q ) η , η ] = 0
η = x , y ( ϵ k η , η Λ η , η ) e k ( η ) ( q ) ν η ( E k , + ( η ) ( 0 ) η , η E k , + ( η ) ( 0 ) η , η ) = 0
η , η = ψ v | j η | ψ c ψ c | j η | ψ v n v ( q )
Λ η , η = ψ v | j η | ψ c ψ c | j η | ψ v ( ϵ μ + ϵ c ϵ v ) n v ( q ) .
2 z 2 E k , + ( η ) ( z ) + i ω μ 0 δ ( z ) η = x , y σ η , η E k , + ( η ) ( z ) + ω 2 ε r μ 0 E k , + ( η ) ( z ) = 0 ; ω k = ϵ k ϵ μ ,
σ η , η = ( i 4 π ϵ r c 2 A 0 ω k ) [ x , x x , y     y , x y , y ] [ ϵ k x , x Λ x , x ϵ k x , y Λ x , y     ϵ k y , x Λ y , x ϵ k y , y Λ y , y ] 1 [ x , x x , y     y , x y , y ] .
σ η , η = ( π ϵ r c 2 A 0 ) ψ v | j η | ψ c ψ c | j η | ψ v i ω k ( ω k ( ϵ c ϵ v ) ) n v ( q ) .
( E k , + ( x , i ) ( 0 ) , E k , + ( y , i ) ( 0 ) ) + ( E k , + ( x , r ) ( 0 ) , E k , + ( y , r ) ( 0 ) ) ) = ( E k , + ( x , t ) ( 0 ) , E k , + ( y , t ) ( 0 ) )
z E k , + ( η , t ) ( 0 ) z E k , + ( η , i ) ( 0 ) z E k , + ( η , r ) ( 0 ) = i ω μ 0 η = x , y σ η , η E k , + ( η ) ( 0 )
[ E k , + ( x , t ) ( 0 )   E k , + ( y , t ) ( 0 ) ] = [ ( 1 + Γ x x ) Γ x y     Γ y x ( 1 + Γ y y ) ] 1 [ E k , + ( x , i ) ( 0 )   E k , + ( y , i ) ( 0 ) ] ,
[ E k , + ( x , r ) ( 0 )   E k , + ( y , r ) ( 0 ) ] = [ ( 1 + Γ x x ) Γ x y     Γ y x ( 1 + Γ y y ) ] 1 [ Γ x x Γ x y     Γ y x Γ y y ] [ E k , + ( x , i ) ( 0 )   E k , + ( y , i ) ( 0 ) ] .
r ± = | r ± | e i θ K ( ± ) = | ( E k , + ( x , r ) ± i E k , + ( y , r ) ) / E k , + ( x , i ) | = ± i G Π x ( Π x ± i Π y ) 2 ω k ( ω k ( ϵ c ϵ v ) + i G ( | Π x | 2 + | Π y | 2 )
t ± = | t ± | e i θ F ( ± ) = | ( E k , + ( x , t ) ± i E k , + ( y , t ) ) / E k , + ( x , i ) | = ω k ( ω k ( ϵ c ϵ v ) ± G Π y ( Π x ± i Π y ) 2 ω k ( ω k ( ϵ c ϵ v ) + i G ( | i x | 2 + | Π y | 2 )
Π η = ± i ( v 1 η ) 2 + ( v 2 η ) 2 s i n ( ϕ q ϕ q , η ( s ) ) , ϕ q , η ( s , τ ) = t a n 1 ( v 2 η / v 1 η ) .
| t y t x | = | 2 ( ω k ( ω k ϵ c v ) ( c μ 0 / A 0 ) Π y Π y ( c μ 0 / A 0 ) Π y Π x | , ϵ c v = ϵ c ϵ v .
T ( ω ) = 1 2 ( | t + | 2 + | t | 2 ) ,    R ( ω ) = 1 2 ( | r + | 2 + | r | 2 )
θ F ( ω ) = 1 2 ( θ F ( + ) θ F ( ) ) ,    θ K ( ω ) = 1 2 ( θ K ( + ) θ K ( ) ) .
S ( 1 ) = η d k t η b R , k η | 0 , μ 0 , μ | b R , k η + d k r η b L , k η | 0 , μ 0 , μ | b L , k η .
g η , η ( 1 ) ( z 1 , z 2 ) = o u t | b R η ( z 1 ) b R η ( z 2 ) | o u t o u t | b R η ( z ) b R η ( z ) | o u t .
g η , η ( 1 ) c l ( z 1 , z 2 ) = d k E k , + ( η ) ( z 2 ) E k , + ( η ) * ( z 1 ) d k E k , + ( η ) ( z ) E k , + ( η ) * ( z ) = d k t η ( k ) t η * ( k ) e i k ( z 1 z 2 ) d k t η ( k ) t η * ( k ) .
g η , η ( 1 ) ( z 1 , z 2 ) = i n | S ( 1 ) b R η ( z 1 ) b R η ( z 2 ) S ( 1 ) | i n + n > 1 i n | S ( n ) b R η ( z 1 ) b R η ( z 2 ) S ( n ) | i n .
ω c ( ± ) = 1 2 ( ϵ c v ± ( ϵ c v ) 2 4     G   ( Π x / Π y ) (   | Π y | 2 + | Π x | 2 ) ) .
ε i s = ( ε 0 0 ε ) ,   ε s s = ( 0 ε ε 0 ) ,   ε a s = ( ε 0 0 ν ε ) ,   ε z s = ( ν ε 0 0 ε ) .
Select as filters


Select Topics Cancel
© Copyright 2024 | Optica Publishing Group. All rights reserved, including rights for text and data mining and training of artificial technologies or similar technologies.