betaFeedback

WorksPhysics and Astronomy

On Parametric Amplification In Discrete Josephson Transmission Line

Abstract

We consider the discrete series-connected lossy Josephson transmission line, constructed from Josephson junctions, capacitors and resistors (one-dimensional array of Josephson junctions). We derive equations describing pump, signal, and idler interaction in the system and calculate the thresholds for the parametric amplification.

Introduction

Superconducting parametric amplifiers attract a lot of interest, due to their importance in microwave electronics [1, 2, 3, 4]. Traditional amplifiers comprise a single Josephson junction (JJ) or an array of junctions in a resonant cavity which ultimately limits the bandwidth and dynamic range.

Recently, owing to impact of the kinetic-inductance traveling-wave parametric amplifier, the Josephson traveling-wave parametric amplifiers enabling larger gain per unit length with lesser pump power have been in the particular focus of several research groups [5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22].

We studied previously [23] the problem of parametric amplification in the Josephson transmission line (JTL) in the continuum approximation. Now we consider the problem for the discrete JTL.

RSJ Josephson transmission line

We consider a model of the JTL presented in Figure 1.

Josephson transmission line composed of superconducting grains

Figure 1. Josephson transmission line composed of superconducting grains

We take as the dynamical variables Josephson phases and the potentials of the grains VnV_n. The circuit equations are [23]

ℏ2edφndt=Vn−1−Vn(1)\frac{\hbar}{2e}\frac{d\varphi_n}{dt}=V_{n-1}-V_n \tag*{(1)}
CdVndt+GVn=∂n2(VnRJ+CJdVndt)+Ic[sin⁡(φn+φ0)−sin⁡(φn+1+φ0)](2)C\frac{dV_n}{dt}+GV_n=\partial_n^2\left(\frac{V_n}{R_J}+C_J\frac{dV_n}{dt}\right)+I_c\left[\sin(\varphi_n+\varphi_0)-\sin(\varphi_{n+1}+\varphi_0)\right] \tag*{(2)}

where CC is the ground capacitance, IcI_c is the critical current of the JJ, GG is the conductance of the ohmic resistor shunting the ground capacitor, RJR_J is the ohmic resistor shunting the JJ; the second-order discrete differential operator ∂n2\partial_n^2 is defined as ∂n2fn=fn−1−2fn+fn+1\partial_n^2 f_n=f_{n-1}-2f_n+f_{n+1}. For the sake of generality we consider d.c. background Josephson current Icsin⁡(φ0)I_c\sin(\varphi_0) flowing along the JTL, where φ0\varphi_0 is the d.c. Josephson phase.

Inclusion of the capacitance in parallel to the Josephson elements (inter-island capacitance) would certainly make the model more physically realistic. However it complicates the mathematics (there appears the fourth order derivative in the final equation). In spite of this the parallel capacitance was included in our previous paper dealing with parametric amplification in the JTL in the continuum approximation [23]. It turned out that such generalization of the model doesn’t lead to any qualitative differences. So in the present paper we decided to give the priority to simplicity over generality.

We can exclude VnV_n and obtain closed equation for φn\varphi_n

d2φndτ2+ZJGdφndτ=∂n2[sin⁡(φn+φ0)+ZJRJdφndτ+CJCd2φndτ2](3)\frac{d^2\varphi_n}{d\tau^2}+Z_JG\frac{d\varphi_n}{d\tau}=\partial_n^2\left[\sin(\varphi_n+\varphi_0)+\frac{Z_J}{R_J}\frac{d\varphi_n}{d\tau}+\frac{C_J}{C}\frac{d^2\varphi_n}{d\tau^2}\right] \tag*{(3)}

where we introduced the dimensionless time τ=t/LJC\tau=t/\sqrt{L_JC}, ZJ≡LJ/CZ_J\equiv\sqrt{L_J/C} is the characteristic impedance of the JTL, and LJ=ℏ/(2eIc)L_J=\hbar/(2eI_c). Let us present (2) as

d2φndτ2−∂n2(cos⁡φ0⋅φn+CJCd2φndτ2)=∂n2{ZJRJdφndτ+n.l.[sin⁡(φn+φ0)]}−ZJGdφndτ(4)\frac{d^2\varphi_n}{d\tau^2}-\partial_n^2\left(\cos\varphi_0\cdot\varphi_n+\frac{C_J}{C}\frac{d^2\varphi_n}{d\tau^2}\right) =\partial_n^2\left\{\frac{Z_J}{R_J}\frac{d\varphi_n}{d\tau}+n.l.[\sin(\varphi_n+\varphi_0)]\right\}-Z_JG\frac{d\varphi_n}{d\tau} \tag*{(4)}

where n.l. stands for the non-linear terms of expansion of the sine function in Taylor series.

If we ignore the dissipation and the non-linear terms, Eq. (4) has obvious solution

φn=12∑αAαei(kαn−ωατ)+c.c.(5)\varphi_n = \frac{1}{2}\sum_{\alpha} A_\alpha e^{i(k_\alpha n-\omega_\alpha\tau)}+\mathrm{c.c.} \tag*{(5)}

where

ωα2=4cos⁡φ0sin⁡2(kα2)1+4CJCsin⁡2(kα2),(6)\omega_\alpha^2=\frac{4\cos\varphi_0\sin^2\left(\frac{k_\alpha}{2}\right)}{1+\frac{4C_J}{C}\sin^2\left(\frac{k_\alpha}{2}\right)}, \tag*{(6)}

and AαA_\alpha are arbitrary constant amplitudes. The nonlinearity and the dissipation we’ll take into account approximately, by changing (4) to

φn=12∑αAα(n)ei(kαn−ωατ)+c.c.(7)\varphi_n=\frac{1}{2}\sum_{\alpha} A_\alpha(n)e^{i(k_\alpha n-\omega_\alpha\tau)}+\mathrm{c.c.} \tag*{(7)}

and assuming that the complex amplitudes Aα(n)A_\alpha(n) slowly change with nn. Additionally, because we consider the dissipation and nonlinear terms as being in some sense small, while calculating the discrete second order derivative of the terms in the parenthesis in (3) we will ignore the corrections which come from the nn-dependence of the amplitudes, thus the discrete second derivative operator acting on a partial wave with the wave vector kk just multiplies this partial wave by −4sin⁡2(k2)-4\sin^2\left(\frac{k}{2}\right). For example,

ddτ(φn−1−2φn−φn+1)=−2i∑αsin⁡2(kα2)ωαAαei(kαn−ωατ)+c.c.,(8)\frac{d}{d\tau}\left(\varphi_{n-1}-2\varphi_n-\varphi_{n+1}\right) =-2i\sum_{\alpha}\sin^2\left(\frac{k_\alpha}{2}\right)\omega_\alpha A_\alpha e^{i(k_\alpha n-\omega_\alpha\tau)}+\mathrm{c.c.}, \tag*{(8)}

However, we’ll do better while calculating the discrete second derivative standing in the l.h.s. of (3). We promote nn, as argument of the amplitudes, to the continuous variable ZZ and approximate

Aα(n±1)=Aα(Z)±∂Aα(Z)∂Z.(9)A_\alpha(n\pm1)=A_\alpha(Z)\pm\frac{\partial A_\alpha(Z)}{\partial Z}. \tag*{(9)}

As the result, for the discrete second derivative we obtain

φn−1−2φn+φn+1=−2∑αsin⁡2(kα2)Aαei(ωατ−kαn)+c.c.+i∑αsin⁡(kα)∂Aα∂Zei(kαn−ωατ)+c.c.(10)\varphi_{n-1}-2\varphi_n+\varphi_{n+1} =-2\sum_{\alpha}\sin^2\left(\frac{k_\alpha}{2}\right)A_\alpha e^{i(\omega_\alpha\tau-k_\alpha n)}+\mathrm{c.c.} +i\sum_{\alpha}\sin(k_\alpha)\frac{\partial A_\alpha}{\partial Z}e^{i(k_\alpha n-\omega_\alpha\tau)}+\mathrm{c.c.} \tag*{(10)}

waves mixing

Coupled equations for the amplitudes

Now let us take into account the nonlinear terms. Consider the case φ0∼1\varphi_0\sim1 and ∣φn∣≪1\lvert\varphi_n\rvert\ll1 Expanding the sine function in the r.h.s. of (2) in Taylor series and keeping the first two terms of the expansion we have

n.l.[sin⁡(φn+φ0)]=−sin⁡φ02φn2.(11)\mathrm{n.l.}\left[\sin\left(\varphi_n+\varphi_0\right)\right]=-\frac{\sin\varphi_0}{2}\varphi_n^2. \tag*{(11)}

After we substitute (4) into (10), the general result for the quadratic term would be too complicated, so we’ll consider a superposition of only three waves (pump, signal, and idler), with the frequencies ωp\omega_p, ωs\omega_s and ωi\omega_i respectively, satisfying equation

ωs+ωi=ωp.(12)\omega_s+\omega_i=\omega_p. \tag*{(12)}

Keeping only the terms which have ωp,s,it\omega_{p,s,i}t time dependence we obtain

φn2=12AsAiei[(ks+ki)n−ωpt]+12ApAi∗ei[(kp−ki)n−ωst]+12ApAs∗ei[(kp−ks)n−ωit]+c.c.(13)\varphi_n^2=\frac{1}{2}A_sA_i e^{i[(k_s+k_i)n-\omega_p t]}+\frac{1}{2}A_pA_i^*e^{i[(k_p-k_i)n-\omega_s t]}+\frac{1}{2}A_pA_s^*e^{i[(k_p-k_s)n-\omega_i t]}+\mathrm{c.c.} \tag*{(13)}

Finally, collective all the perturbative terms in (2) we obtain coupled equations for the amplitudes

dApdZ+νpAp=iμpAsAieiΔk⋅Z(14)\frac{dA_p}{dZ}+\nu_pA_p=i\mu_pA_sA_i e^{i\Delta k\cdot Z} \tag*{(14)}
dAs,idZ+νsAs,i=iμs,iApAi,s∗e−iΔk⋅Z,(15)\frac{dA_{s,i}}{dZ}+\nu_sA_{s,i}=i\mu_{s,i}A_pA_{i,s}^*e^{-i\Delta k\cdot Z}, \tag*{(15)}

where

να=tan⁡(kα/2)ωα[ZJG+4sin⁡2(kα2)ZJRJ](16)\nu_\alpha=\frac{\tan(k_\alpha/2)}{\omega_\alpha}\left[Z_JG+4\sin^2\left(\frac{k_\alpha}{2}\right)\frac{Z_J}{R_J}\right] \tag*{(16)}
μp=2sin⁡φ0⋅tan⁡(kp/2)ωp2sin⁡2(ks+ki2)(17)\mu_p=\frac{2\sin\varphi_0\cdot\tan(k_p/2)}{\omega_p^2}\sin^2\left(\frac{k_s+k_i}{2}\right) \tag*{(17)}
μs,i=2sin⁡φ0⋅tan⁡(ks,i/2)ωs,i2sin⁡2(kp−ki,s2)(18)\mu_{s,i}=\frac{2\sin\varphi_0\cdot\tan(k_{s,i}/2)}{\omega_{s,i}^2}\sin^2\left(\frac{k_{p}-k_{i,s}}{2}\right) \tag*{(18)}
Δk=ks+ki−kp.(19)\Delta k=k_s+k_i-k_p. \tag*{(19)}

Comparing Eqs. (13) with the appropriate equations from our previous publication [23], we see that the basic equations systems in the discrete consideration and in continuum approximation are the same. Only the parameters of the systems are different (compare (14) with the appropriate equations from our previous publication [23]).

Weak signal: threshold for the parametric amplification

Let us solve Eq. (13) in the small (relative to the pump) signal and idler approximation (∣As∣,∣Ai∣≪∣Ap∣\lvert A_s\rvert,\lvert A_i\rvert\ll\lvert A_p\rvert). In this approximation the equation for ApA_p becomes decoupled from the other two and takes the form

dApdZ+νpAp=0,(20)\frac{dA_p}{dZ}+\nu_pA_p=0, \tag*{(20)}

with the obvious solution

Ap(Z)=e−νpZAp0.(21)A_p(Z)=e^{-\nu_p Z}A_p^0. \tag*{(21)}

Treating the other two equations, we’ll introduce the local approximation, by treating ApA_p in the r.h.s. of (15) as a constant. In this approximation the solutions for As,AiA_s,A_i are

\begin{pmatrix} A_s \\ A_i \end{pmatrix} =e^{-i\Delta k/2\cdot Z} \begin{pmatrix} \cdots e^{KZ} \\ \cdots e^{K^*Z} \end{pmatrix}, \tag*{(22)}\end{pmatrix}

where dots stand for some constant amplitudes, and KK is one of the roots of the characteristic polynomial

P(K)≡K2+(νs+νi)K+(νs−iΔk/2)(νi+iΔk/2)−μsμi∣Ap∣2.(23)P(K)\equiv K^2+(\nu_s+\nu_i)K+(\nu_s-i\Delta k/2)(\nu_i+i\Delta k/2)-\mu_s\mu_i\lvert A_p\rvert^2. \tag*{(23)}

Parametric amplification of the signal takes place when one of the roots have positive real part. The boundary between the parametric amplification and no parametric amplification we can find by demanding that KK for that boundary is purely imaginary. So, after simple algebra, we obtain the condition for the parametric amplification:

μsμi∣Ap∣2>νsνp[1+(Δk)2(νs+νi)2].(24)\mu_s\mu_i \lvert A_p\rvert^2 > \nu_s\nu_p\left[1+\frac{(\Delta k)^2}{(\nu_s+\nu_i)^2}\right]. \tag*{(24)}

Looking at Eq. (24) we realise that the momenta mismatch Δk\Delta k acts in some sense similar to the losses in the system. Both factors together define the threshold for the parametric amplification.

Though it’s not completely consistent, we bring back the coordinate dependence of ApA_p, given by Eq. (21), and consider (24) as a local condition for the parametric amplification.

Small wave vectors limiting case

Consider the limiting case kp,ks,ki≪1k_p,k_s,k_i \ll1. In this case Δk=0\Delta k=0. We assume additionally that RJ=∞R_J=\infty, hence να=ν\nu_\alpha=\nu (doesn’t depend upon α\alpha). Hence Eqs. (13a), (13b) take the form

dApdZ+νAp=iμpAsAi(25)\frac{dA_p}{dZ}+\nu A_p=i\mu_p A_sA_i \tag*{(25)}
dAs,idZ+νAs,i=iμs,iApAi,s∗.(26)\frac{dA_{s,i}}{dZ}+\nu A_{s,i}=i\mu_{s,i}A_pA_{i,s}^{*}. \tag*{(26)}

where

μp=μs+μi.(27)\mu_p=\mu_s+\mu_i. \tag*{(27)}

We’ll present the amplitudes as

Ap=ie−νZAp(r),As,i=e−νZAs,i(r).(28)A_p=ie^{-\nu Z}A_p^{(r)},\qquad A_{s,i}=e^{-\nu Z}A_{s,i}^{(r)}. \tag*{(28)}

In the new variables (25), (20b) take the form (compare with Ref. [24])

dAp(r)dZ~=μpAs(r)Ai(r)(29)\frac{dA_p^{(r)}}{d\widetilde{Z}}=\mu_p A_s^{(r)}A_i^{(r)} \tag*{(29)}
dAs,i(r)dZ~=−μs,iAp(r)Ai,s(r).(30)\frac{dA_{s,i}^{(r)}}{d\widetilde{Z}}=-\mu_{s,i}A_p^{(r)}A_{i,s}^{(r)}. \tag*{(30)}

where

dZ~≡e−νZdZ,(31)d\widetilde{Z}\equiv e^{-\nu Z}dZ, \tag*{(31)}

that is

Z~=K−1νe−νZ,(32)\widetilde{Z}=K-\frac{1}{\nu}e^{-\nu Z}, \tag*{(32)}

where KK is an arbitrary constant.

We’ll consider only the real solutions, postponing the general analysis until later. From (23a), (23b) follows

12ddZ~[(Ap(r))2+(As(r))2+(Ai(r))2]=(μp−μs−μi)Ap(r)As(r)Ai(r).(33)\frac{1}{2}\frac{d}{d\widetilde{Z}}\left[(A_p^{(r)})^2+(A_s^{(r)})^2+(A_i^{(r)})^2\right] =(\mu_p-\mu_s-\mu_i)A_p^{(r)}A_s^{(r)}A_i^{(r)}. \tag*{(33)}

Due to (21) the r.h.s. of (26) is equal to zero, hence

(Ap(r))2+(As(r))2+(Ai(r))2=const,(34)(A_p^{(r)})^2+(A_s^{(r)})^2+(A_i^{(r)})^2=\mathrm{const}, \tag*{(34)}

and we can look for a solution in the form (remember the spherical coordinates)

Ap(r)=Bcos⁡θ(35)A_p^{(r)}=B\cos\theta \tag*{(35)}
As(r)=Bsin⁡θcos⁡φ(36)A_s^{(r)}=B\sin\theta\cos\varphi \tag*{(36)}
Ai(r)=Bsin⁡θsin⁡φ,(37)A_i^{(r)}=B\sin\theta\sin\varphi, \tag*{(37)}

where BB is an arbitrary amplitude. For the new variables θ\theta and φ\varphi we obtain equations

dθdZ~=−B2μpsin⁡θsin⁡2φ(38)\frac{d\theta}{d\widetilde{Z}}=-\frac{B}{2}\mu_p\sin\theta\sin2\varphi \tag*{(38)}
dφdZ~=−B2cos⁡θ(μpcos⁡2φ−μs+μi).(39)\frac{d\varphi}{d\widetilde{Z}}=-\frac{B}{2}\cos\theta(\mu_p\cos2\varphi-\mu_s+\mu_i). \tag*{(39)}

Dividing one of the equations to the other we obtain

dθtan⁡θ=μpsin⁡2φ dφμpcos⁡2φ−μs+μi.(40)\frac{d\theta}{\tan\theta}=\frac{\mu_p\sin2\varphi\,d\varphi}{\mu_p\cos2\varphi-\mu_s+\mu_i}. \tag*{(40)}

(40) can be easily integrated and we obtain

sin⁡2θ=Dμpcos⁡2φ−μs+μi,(41)\sin^2\theta=\frac{D}{\mu_p\cos2\varphi-\mu_s+\mu_i}, \tag*{(41)}

where DD is the integration constant. The solution (31), plotted in the variables Aα(r)A_\alpha^{(r)}, is graphically presented in Figure 2. In the variables AαA_\alpha the sphere contracts uniformly.

The solution (31) plotted in the variables $A_\alpha^{(r)}$

Figure 2. The solution (31) plotted in the variables Aα(r)A_\alpha^{(r)}. We have chosen μi/μs=1.2\mu_i/\mu_s=1.2 and θ=.1\theta=.1 for φ=0\varphi=0.

(41) is even simpler than it looks. Taking into account (27), it can be written down as

(μs−μi)(Ap(r))2+μp((As(r))2−(Ai(r))2)=const.(42)(\mu_s-\mu_i)(A_p^{(r)})^2+\mu_p\left((A_s^{(r)})^2-(A_i^{(r)})^2\right)=\mathrm{const}. \tag*{(42)}

The latter equation follows immediately from (23a), (23b). There also exists a singular solution of (29a), (29b), where the constant φ\varphi is given by the equation

μpcos⁡2φ−μs+μi=0,(43)\mu_p\cos2\varphi-\mu_s+\mu_i=0, \tag*{(43)}

and θ(Z~)\theta(\widetilde{Z}), found from (29a), is given by the equation

cos⁡θ=tanh⁡(B2μpsin⁡2ϕ⋅Z~).(44)\cos\theta=\tanh\left(\frac{B}{2}\mu_p\sin2\phi\cdot\widetilde{Z}\right). \tag*{(44)}

(44) shows that when Z~→∞\widetilde{Z}\to\infty, either θ→0\theta\to0 or θ→π\theta\to\pi, depending upon the sign of sin⁡2ϕ\sin2\phi.

However, we are mostly interested not in the dependence of one angle variable upon the other, but in their dependence upon Z~\widetilde{Z}. This dependence can be found by numerical integration of Eqs. (29a), (29b) and is illustrated in Figure 3. The left side of the Figure, with θ\theta increasing from nearly 0, to π/2\pi/2, demonstrates the parametric amplification (provided the losses are not too big).

Figure 3. Numerical solution of Eqs. (29a), (29b) for θ\theta (blue solid line) and ϕ\phi (red dashed line). We have chosen μs=.2\mu_s=.2, μi=.24\mu_i=.24, θ(0)=.1\theta(0)=.1 and ϕ(0)=0\phi(0)=0 (no idle signal at Z~=0\widetilde{Z}=0).

waves mixing

Coupled equations for the amplitudes

Now let us consider the case ϕ0=0\phi_0=0. In this case, expanding the sine function in the r.h.s. of (2) in Taylor series with respect to phases difference and keeping the first two nonzero terms of the expansion we have

n.l.[sin⁡(ϕn)]=−16ϕn3.(45)n.l.[\sin(\phi_n)]=-\frac{1}{6}\phi_n^3. \tag*{(45)}

We again consider the superposition of three waves, but the frequencies this time satisfy the relation

ωs+ωi=2ωp.(46)\omega_s+\omega_i=2\omega_p. \tag*{(46)}

Keeping only the terms which have ωp,s,it\omega_{p,s,i}t time dependence we obtain

ϕn3=38∑α,β(2−δαβ)Aα∣Aβ∣2ei[kαn−ωαt]+34Ap∗AsAiei[(ks+ki−kp)n−ωpt]+38Ap2Ai∗ei[(2kp−ki)n−ωst]+38Ap2As∗ei[(2kp−ks)n−ωit]+c.c.(47)\begin{aligned} \phi_n^3={}&\frac{3}{8}\sum_{\alpha,\beta}(2-\delta_{\alpha\beta})A_\alpha\lvert A_\beta\rvert^2e^{i[k_\alpha n-\omega_\alpha t]}\\ &+\frac{3}{4}A_p^*A_sA_ie^{i[(k_s+k_i-k_p)n-\omega_p t]}+\frac{3}{8}A_p^2A_i^*e^{i[(2k_p-k_i)n-\omega_s t]}\\ &+\frac{3}{8}A_p^2A_s^*e^{i[(2k_p-k_s)n-\omega_i t]}+\text{c.c.} \tag*{(47)} \end{aligned}

Finally, collective all the perturbative terms in (2) we obtain coupled equations for the amplitudes

dApdZ+νpAp=iγp(∣Ap∣2+2∣As∣2+2∣Ai∣2)Ap+2iηpAp∗AsAieiΔk⋅Z(48)\frac{dA_p}{dZ}+\nu_p A_p=i\gamma_p\left(\lvert A_p\rvert^2+2\lvert A_s\rvert^2+2\lvert A_i\rvert^2\right)A_p+2i\eta_p A_p^*A_sA_i e^{i\Delta k\cdot Z} \tag*{(48)}
dAs,idZ+νs,iAs,i=iγs,i(2∣Ap∣2+∣As,i∣2+2∣Ai,s∣2)As,i+iηs,iAp2Ai,s∗e−iΔk⋅Z,(49)\frac{dA_{s,i}}{dZ}+\nu_{s,i}A_{s,i}=i\gamma_{s,i}\left(2\lvert A_p\rvert^2+\lvert A_{s,i}\rvert^2+2\lvert A_{i,s}\rvert^2\right)A_{s,i}+i\eta_{s,i}A_p^2A_{i,s}^*e^{-i\Delta k\cdot Z}, \tag*{(49)}

where να\nu_\alpha is given by (14a) with cos⁡ϕ0\cos\phi_0 being put equal to 1, and

γα=tan⁡(kα/2)2ωp2sin⁡2(kα2)(50)\gamma_\alpha=\frac{\tan(k_\alpha/2)}{2\omega_p^2}\sin^2\left(\frac{k_\alpha}{2}\right) \tag*{(50)}
ηp=tan⁡(kp/2)2ωp2sin⁡2(ks+ki−kp2)(51)\eta_p=\frac{\tan(k_p/2)}{2\omega_p^2}\sin^2\left(\frac{k_s+k_i-k_p}{2}\right) \tag*{(51)}
ηs,i=tan⁡(ks,i/2)2ωp2sin⁡2(2kp−ki,s2)(52)\eta_{s,i}=\frac{\tan(k_{s,i}/2)}{2\omega_p^2}\sin^2\left(\frac{2k_p-k_{i,s}}{2}\right) \tag*{(52)}
Δk=ks+ki−2kp(53)\Delta k=k_s+k_i-2k_p \tag*{(53)}

Comparing Eqs. (38) with the appropriate equations from our previous publication [23] we can only repeat what was said in the very end of § 3-A.

Weak signal

In the small (relative to the pump) signal and idler approximation (∣As∣,∣Ai∣≪∣Ap∣\lvert A_s\rvert,\lvert A_i\rvert\ll\lvert A_p\rvert), the equation for ApA_p becomes decoupled from the other two and takes the form

dApdZ+νpAp=i2γp∣Ap∣2Ap.(54)\frac{dA_p}{dZ}+\nu_p A_p=\frac{i}{2}\gamma_p\lvert A_p\rvert^2A_p. \tag*{(54)}

It is interesting to compare the result following from equation (40) with the nonlinear dispersion law obtained in our previous publication [25]. There we considered equation

d2ϕndτ2=sin⁡ϕn−1−2sin⁡ϕn+sin⁡ϕn+1(55)\frac{d^2\phi_n}{d\tau^2}=\sin\phi_{n-1}-2\sin\phi_n+\sin\phi_{n+1} \tag*{(55)}

and obtained in the up to the second order with respect to the amplitude approximation the solution

ϕn(τ)=acos⁡[kn−ω(k)τ](56)\phi_n(\tau)=a\cos[kn-\omega(k)\tau] \tag*{(56)}

with the dispersion law

ω(k)=2(1−a216)sin⁡(k2).(57)\omega(k)=2\left(1-\frac{a^2}{16}\right)\sin\left(\frac{k}{2}\right). \tag*{(57)}

In the absence of dissipation, the solution of (40) can be taken as

Ap(Z)=aeiγpa2/2Z,(58)A_p(Z)=ae^{i\gamma_p a^2/2 Z}, \tag*{(58)}

where aa is a real constant. Substituting this solution into (6) and taking into account (39a) we obtain

ϕn(τ)=acos⁡[(k+a2tan⁡(k/2)/8)n−ωτ],(59)\phi_n(\tau)=a\cos\left[(k+a^2\tan(k/2)/8)n-\omega\tau\right], \tag*{(59)}

where ω=2sin⁡(k/2)\omega=2\sin(k/2). let us introduce

k′=k+a2tan⁡(k/2)/8.(60)k'=k+a^2\tan(k/2)/8. \tag*{(60)}

Then up to the lowest order with respect to a2a^2 corrections

k=k′−a2tan⁡(k′/2)8,(61)k = k' - \frac{a^2 \tan(k'/2)}{8}, \tag*{(61)}

and the dispersion law can be presented as

ω=2sin⁡[k′/2−a2tan⁡(k′/2)16].(62)\omega= 2 \sin\left[k'/2 - \frac{a^2 \tan(k'/2)}{16}\right]. \tag*{(62)}

Expanding (48) in series with respect to a2a^2 and keeping the terms up to the linear order with respect to a2a^2 we obtain

ω(k′)=2(1−a216)sin⁡(k′2).(63)\omega(k') = 2\left(1-\frac{a^2}{16}\right)\sin\left(\frac{k'}{2}\right). \tag*{(63)}

which coincides with (57).

Note that in our previous publications [25] the pump wave hire harmonics, which exist simultaneously with the main one, due to nonlinearity of the system, are also considered. However there influence on the signal amplitude would be an effect of the higher order with respect to the parameter Ap2A_p^2, than the influence of the main harmonic.

Threshold for the parametric amplification

(49) in the weak signal approximation can be written down as

dAs,idZ+(νs,i−iγs,i∣Ap∣2)As,i=iηs,iAp2Ai,s∗e−iΔk⋅Z.(64)\frac{dA_{s,i}}{dZ} + \left(\nu_{s,i} - i\gamma_{s,i}|A_p|^2\right) A_{s,i} = i\eta_{s,i}A_p^2 A_{i,s}^{*}e^{-i\Delta k\cdot Z}. \tag*{(64)}

In the framework of the local approximation, we treat ApA_p in the r.h.s. of Eqs. (50) as a constant, so we recover (17), where KK this time is one of the roots of the characteristic polynomial

P(K)≡K2+(νs+νi)K+(νs−2iγs∣Ap∣2−iΔk/2)(νi+2iγi∣Ap∣2+iΔk/2)−ηsηi∣Ap∣4.(65)P(K) \equiv K^2 + (\nu_s+\nu_i)K + \left(\nu_s - 2i\gamma_s|A_p|^2 - i\Delta k/2\right)\left(\nu_i + 2i\gamma_i|A_p|^2 + i\Delta k/2\right) - \eta_s\eta_i|A_p|^4. \tag*{(65)}

We again find the boundary between the parametric amplification and no amplification by demanding that KK for that boundary is purely imaginary. So, after simple algebra, we obtain the condition for the parametric amplification:

ηsηi∣Ap∣4>νsνi{1+[Δk+2(γs+γi)∣Ap∣2]2(νs+νi)2}.(66)\eta_s\eta_i|A_p|^4 > \nu_s\nu_i\left\{1+\frac{\left[\Delta k+2(\gamma_s+\gamma_i)|A_p|^2\right]^2}{(\nu_s+\nu_i)^2}\right\}. \tag*{(66)}

Here we can only repeat what was said after Eq. (19). Again, we bring back the coordinate dependence of ApA_p, given by Eq. (40) and consider (52) as a local condition for the parametric amplification.

Discussion

The present paper generalized the previously obtained results on parametric amplification in JTL [5, 11, 25] to the discrete model of the line. The discrete character of the line leads to the renormalization of the parameters of the coupled amplitude equations.

We formulate the local approximation which allows explicitly find the threshold for parametric amplification in the case of weak signal. In addition we present the exact particular solution for the coupled equations for the pump, signal and idler amplitudes in the particular case of small wave vectors and present the solution graphically. Finally, we would like to mention that in real case of the line of finite length, the frequency dependence of the line impedance leads to matching problems and hence to the wave reflection (e.g., [26]). However we leave the study of the problem until later.

Acknowledgment

I am grateful to T. H. A. van der Reep for the discussion. I am also very grateful to Donostia International Physics Center (DIPC) for the hospitality during my visit, when the paper was finalised.

References

  1. [1]M. A. Castellanos-Beltran and K. W. Lehnert, Appl. Phys. Lett. 91, 083509 (2007).
  2. [2]P. D. Nation, J. R. Johansson, M. P. Blencowe, and F. Nori, Rev. Mod. Phys. 84, 1 (2012).
  3. [3]B. H. Eom, P. K. Day, H. G. LeDuc, and J. Zmuidzinas, Nat. Phys. 8, 623 (2012).
  4. [4]J. Aumentado, IEEE Microwave magazine, 21, 45 (2020).
  5. [5]O. Yaakobi, L. Friedland, C. Macklin, and I. Siddiqi, Phys. Rev. B 87, 144301 (2013).
  6. [6]K. O’Brien, C. Macklin, I. Siddiqi, and X. Zhang, Phys. Rev. Lett. 113, 157001 (2014).
  7. [7]C. Macklin, K. O’Brien, D. Hover, M. E. Schwartz, V. Bolkhovsky, X. Zhang, W. D. Oliver, and I. Siddiqi, Science 350, 307 (2015).
  8. [8]B. A. Kochetov, and A. Federov, Phys. Rev. B. 92, 224304 (2015).
  9. [9]M.T. Bell and A. Samolov, Phys. Rev. Applied 4, 024014 (2015).
  10. [10]T. C. White et al., Appl. Phys. Lett. 106, 242601 (2015).
  11. [11]A. B. Zorin, Phys. Rev. Applied 6, 034006 (2016).
  12. [12]L. Fasolo, A. Greco, E. Enrico, J. Thirumalai, and S. I. Pokutnyi, Adv. Condensed - Matter Materials Physics - Rudimentary Res. Topical Technol. IntechOpen, 2019.
  13. [13]D. M. Basko, F. Pfeiffer, P. Adamus, M. Holzmann, and F. W. J. Hekking, Phys. Rev. B 101, 024518 (2020).
  14. [14]T. Dixon, J. W. Dunstan, G. B. Long, J. M. Williams, Ph. J. Meeson, C. D. Shelly, Phys. Rev. Applied 14, 034058 (2020)
  15. [15]A. B. Zorin, Phys. Rev. Applied 12, 044051 (2019).
  16. [16]A. Miano and O. A. Mukhanov, IEEE Trans. Appl. Supercond. 29, 1501706 (2019).
  17. [17]T. H. A. van der Reep, Phys. Rev. A 99, 063838 (2019).
  18. [18]Ch. Liu, Tzu-Chiao Chien, M. Hatridge, D. Pekker, Phys. Rev. A 101, 042323 (2020).
  19. [19]A. Greco, L. Fasolo, A. Meda, L. Callegaro,and E. Enrico, Phys. Rev. B, 104, 184517 (2021).
  20. [20]Q. Yuan, M. Haider, J. A. Russer, P. Russer, and Ch. Jirausechek. In European Quantum Electronics Conference, 22, Optical Society of America (2021).
  21. [21]C. Kow, V. Podolsky, and A. Kamal, arXiv:2201.04660.arxiv.org/abs/2201.04660
  22. [22]H. Katayama, N. Hatanekaka, T. Fujii, and M. P. Blencowe, Phys. Rev. Res. 5, L022055 (2023).
  23. [23]E. Kogan, Phys. Stat. Sol. (b), Vol. 260, p. 2300005, 2023.
  24. [24]A. L. Cullen, Proceedings of the IEE-Part B: Electronic and Communication Engineering 107, 101 (1960).
  25. [25]E. Kogan, Phys. Stat. Sol. (b), Vol. 259, p. 2200160, 2022.
  26. [26]S. Kern, P. Neilinger, E. Il’ichev, et al., arXiv: 2023.02448v2 (2023).

Paper details

vlnc:W_827n8yj5
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