1: In order to complement the analysis of the $t$-$J$ model,
2: in this section we consider the weakly interacting Hubbard model on a two-leg
3: ladder.
4: A renormalization group (RG) treatment supplemented by Abelian
5: bosonization allows a detailed analysis of the phase diagram in the weakly
6: interacting limit,
7: and a characterization of the various phases in terms of the low-energy
8: modes. We follow an approach established on standard two- and N-leg ladder systems (i.e. for $\Delta
9: \mu=0$).\cite{balents,lin,linso8,ledermann}
10:
11: In the current case of an inhomogeneously doped two-leg ladder we consider the
12: weak repulsive limit, $0<U \ll t,t'$ of the Hubbard model,
13: %\begin{subequations}
14: \begin{eqnarray}\label{hubbardham}
15: H & = & - t \sum_{j,a,s} ( c^{\dag}_{j,a,s} c_{j+1,a,s} +\mathrm{h.c.} )\nonumber \\
16: & & - t' \sum_{j,s} ( c^{\dag}_{j,1,s}
17: c_{j,2,s} + \mathrm{h.c.} ) \\
18: && + U \sum_{j,a} c^{\dag}_{j,a,\uparrow} c_{j,a,\uparrow}
19: c^{\dag}_{j,a,\downarrow} c_{j,a,\downarrow} \nonumber \\
20: && - \sum_{j,a,s} \mu_a c^{\dag}_{j,a,s} c_{j,a,s}, \nonumber
21: \end{eqnarray}
22: %\end{subequations}
23: with the same notations as is Sec. I. The quadratic part of the Hamiltonian
24: (\ref{hubbardham}), i.e. Eq. (\ref{hubbardham}) for $U=0$, can be
25: decoupled via a canonical transformation
26: \begin{equation}
27: d^{\dagger}_{j,1/2,s}=\sqrt{\frac{1\mp\Delta\mu/D}{2}}
28: c^{\dagger}_{j,2,s}\pm \sqrt{\frac{1\pm\Delta\mu/D}{2}} c^{\dagger}_{j,1,s} ,
29: \end{equation}
30: where $D=\sqrt{4{t'}^2+\Delta\mu^2}$. These rung operators interpolate smoothly
31: between the bonding and anti-bonding combinations at $\Delta\mu=0$, and the
32: original fermions for $\Delta\mu/t'\rightarrow \infty$, where
33: $d^{\dagger}_{j,i,s}\rightarrow c^{\dagger}_{j,i,s}$.
34: In momentum space two bands
35: corresponding to
36: $d^{\dagger}_{i,s}(k)$, $i=1,2$
37: result with dispersions
38: \begin{equation}\label{hoppingdisperions}
39: \varepsilon_{1/2}(k)=-2t \cos(k) \pm \frac{1}{2}D-\bar{\mu},
40: \end{equation}
41: and a bandwidth $4t$.
42:
43: Consider now the effect of the Hubbard interaction term
44: in Eq. (\ref{hubbardham}).
45: When both bands are completely separated in energy, only the lower band is filled,
46: and at half filling ($\delta=0$) a band insulator is obtained which upon doping ($\delta>0$) becomes an
47: ordinary spin-1/2 Luttinger liquid (LL).
48: For $D<4t
49: \cos^2(\pi\delta)$, both bands are partially filled, and inter-band interaction effects
50: must be examined. While this proves difficult in general, progress can be made
51: upon considering the weakly interacting limit.
52: Since the interest is thus on the low-energy physics, the
53: dispersions (\ref{hoppingdisperions}) can be linearized around the Fermi points $k_{F,i}$, $i=1,2$, determined
54: by
55: $\varepsilon_1(k_{F,1})$=$\varepsilon_2(k_{F,2})$ and
56: $k_{F,1}+k_{F,2}=\pi(1-\delta)$. Furthermore left- and right
57: movers, $d^{\dagger}_{R/L,i,s}$, are defined with respect to the Fermi level in each band,
58: $i=1,2$. For
59: generic (i.e. incommensurate) Fermi momenta the
60: interactions consist of intra- and inter-band forward- and Cooper- scattering. These
61: can be organized in terms of the U(1) and SU(2) current
62: operators
63: \begin{eqnarray}
64: J_{pij}&=&\sum_s d^{\dagger}_{p,i,s} d_{p,j,s},\nonumber \\
65: J^{\alpha}_{pij}&=&\frac{1}{2}\sum_{s,s'} d^{\dagger}_{p,i,s}
66: \sigma^{\alpha}_{s,s'} d_{p,j,s'} ,
67: \end{eqnarray}
68: where $p=R,L$, and the band indices $i,j=1,2$.
69: The following non-chiral current-current interactions are allowed by symmetry,
70: %
71: \begin{eqnarray}
72: H_{I}&=&\sum_{i}\int\!\! dx\: (c^{\rho}_{ii} J_{Rii} J_{Lii} - c^{\sigma}_{ii}
73: {\bf J}_{Rii}\cdot {\bf J}_{Lii}) \nonumber \\
74: &+& \sum_{i\neq j} \int\!\! dx\: (c^{\rho}_{ij} J_{Rij} J_{Lij} - c^{\sigma}_{ij}
75: {\bf J}_{Rij}\cdot {\bf J}_{Lij})\quad \\
76: &+& \sum_{i\neq j} \int \!\!dx\: (f^{\rho}_{ij} J_{Rii} J_{Ljj} - f^{\sigma}_{ij}
77: {\bf J}_{Rii}\cdot {\bf J}_{Ljj}).\nonumber
78: \end{eqnarray}
79: In this representation $f$ ($c$) denotes couplings related to forward- (Cooper-)
80: scattering, and the
81: symmetry of the inter-band scattering terms under the band exchange is
82: explicitly taken into account.
83: Using current algebra and operator product expansions, a one-loop RG flow for the
84: various couplings can be derived, \cite{balents} which in our notation reads
85: \begin{eqnarray}\label{RGE}
86: \frac{d c^{\rho}_{ii}}{dl} &=& -\frac{1}{2v_{\bar{i}}} \left[
87: (c^{\rho}_{i\bar{i}})^2 +\frac{3}{16}(c^{\sigma}_{i\bar{i}})^2
88: \right],\nonumber \\
89: \frac{d c^{\sigma}_{ii}}{dl} &=& -\frac{1}{2v_{\bar{i}}} \left[
90: \frac{1}{2}(c^{\sigma}_{i\bar{i}})^2 + 2 c^{\rho}_{i\bar{i}}
91: c^{\sigma}_{i\bar{i}} \right]-\frac{1}{2v_i}(c^{\sigma}_{ii})^2,\nonumber \\
92: \frac{d c^{\rho}_{12}}{dl} &=& - \sum_{i}\frac{1}{2 v_i} \left[ c^{\rho}_{1i}
93: c^{\rho}_{i2} + \frac{3}{16} c^{\sigma}_{1i} c^{\sigma}_{i2}
94: \right]\nonumber \\
95: &&+ \frac{2}{v_1 + v_2}\left[ c^{\rho}_{12} f^{\rho}_{12} + \frac{3}{16}
96: c^{\sigma}_{12} f^{\sigma}_{12}\right],\nonumber \\
97: \frac{d c^{\sigma}_{12}}{dl} &=& - \sum_{i}\frac{1}{2 v_i} \left[ c^{\rho}_{1i}
98: c^{\sigma}_{i2} + c^{\sigma}_{1i}
99: c^{\rho}_{i2} + \frac{1}{2} c^{\sigma}_{1i} c^{\sigma}_{i2} \right]
100: \\
101: &&+ \frac{2}{v_1 + v_2}\left[ c^{\rho}_{12} f^{\sigma}_{12} + c^{\sigma}_{12} f^{\rho}_{12} - \frac{1}{2}
102: c^{\sigma}_{12} f^{\sigma}_{12}\right], \nonumber \\
103: \frac{d f^{\rho}_{12}}{dl} &=&\frac{1}{v_1 + v_2}\left[ (c^{\rho}_{12})^2 +
104: \frac{3}{16}(c^{\sigma}_{12})^2\right],\nonumber \\
105: \frac{d f^{\sigma}_{12}}{dl} &=& \frac{1}{v_1 + v_2}\left[ 2 c^{\rho}_{12}
106: c^{\sigma}_{12}-\frac{1}{2}(c^{\sigma}_{12})^2-(f^{\sigma}_{12})^2\right],\nonumber
107: \end{eqnarray}
108: where $i=1,2$, $\bar{1}=2$, $\bar{2}=1$, and $v_{i}=2 t \sin(k_{F,i})$ are the
109: Fermi velocities for the bands. The successive elimination of high frequency
110: modes is obtained from (\ref{RGE}) by integration along the logarithmic length scale $l$, related to
111: an energy scale $E \sim t e^{-\pi l}$ .
112: The flow equations can be integrated once the
113: bare values of the couplings are known. For the Hubbard
114: interaction of Eq. (\ref{hubbardham}) they are obtained as
115: \begin{eqnarray}
116: c^{\sigma}_{11}&=&c^{\sigma}_{22}=4 c^{\rho}_{11}=4 c^{\rho}_{22}=U \left[1+\left(\frac{\Delta\mu}{D}\right)^2\right], \nonumber \\
117: c^{\sigma}_{12}&=&4 c^{\rho}_{12}=f^{\sigma}_{12}=4
118: f^{\rho}_{12}=U \left[1-\left(\frac{\Delta\mu}{D}\right)^2\right].
119: \end{eqnarray}
120: Increasing $\Delta\mu$ away from zero can be seen to reduce the bare inter-band scattering with respect to
121: intra-band scattering.
122:
123: Depending on the parameters, integration of the flow equations leads to
124: different asymptotic behavior, with
125: either a flow
126: to a finite-valued fix point, or to instabilities characterized by universal ratios
127: of the renormalized couplings beyond a scale $l^*$, where the most diverging coupling
128: becomes of the order the bandwidth. While the consistency of the one-loop
129: renormalization group equations is restricted to $l<l^*$, the
130: asymptotic ratios can be utilized to derive a description of the low-energy physics
131: of the system within Abelian bosonization.\cite{haldane} Introducing canonically
132: conjugated bosonic fields $\Phi_{\nu i}$, and $\Pi_{\nu i}$ for the charge and
133: spin degrees of freedom ($\nu=\rho,\sigma$) on each band $i=1,2$, the
134: fermionic operators can be represented as
135: \begin{equation}
136: d_{R/L,i,s}=\frac{\eta_{is}}{\sqrt{2\pi\alpha}} \: \mbox{e}^{\: i
137: \sqrt{\pi/2}[\pm(\Phi_{\rho i} + s \Phi_{\sigma i}) - (\theta_{\rho i}+
138: s\theta_{\sigma i}) ] },
139: \end{equation}
140: where $\theta_{\nu i}$ is the dual field of
141: $\Phi_{\nu i}$, so that $\partial_x \theta_{\nu i}=\Pi_{\nu i}$. The $\eta_{i s}$ are
142: Klein factors, ensuring anticommutation relations, and $\alpha$ is a
143: short-distance cutoff.\cite{schulz} Using the above representation, the interacting fermionic
144: Hamiltonian transforms into a bosonic Hamiltonian,
145: $H_B=H_q+H_{I}$, containing quadratic terms
146: \begin{eqnarray}\label{bosonhamq}
147: H_q&=& \sum_{\nu,i } \frac{1}{2} \int \!\! dx\left[
148: v_i+\frac{c^{\nu}_{ii}}{\beta_{\nu}|\beta_{\nu}| }\right] \partial_x\!
149: \Phi_{\nu i}^2 + \left[v_i-\frac{c^{\nu}_{ii}}{\beta_{\nu}|\beta_{\nu}|}
150: \right] \Pi_{\nu i}^2 \nonumber \\
151: & & + \int \!\! dx\: \frac{f^{\nu}_{12}}{\beta_{\nu}|\beta_{\nu}|} (\partial_x\Phi_{\nu 1}
152: \partial_x\Phi_{\nu 2} - \Pi_{\nu 1}\Pi_{\nu 2}),
153: \end{eqnarray}
154: and sine-Gordon-like interaction terms
155: \begin{eqnarray}\label{bosonhami}
156: H_{I}&=& \int \!\! dx\: \{ \:c^{\sigma}_{11}\cos{(\sqrt{2} \beta_{\sigma} \Phi_{\sigma_1})} +
157: c^{\sigma}_{22}\cos{(\sqrt{2} \beta_{\sigma} \Phi_{\sigma_2})} \nonumber \\
158: & & -4c^{\rho}_{12} \cos{(2 \beta_{\rho}\theta_{\rho -})}
159: [\cos{(\beta_{\sigma} \Phi_{\sigma -})} - \cos{(\beta_{\sigma}\theta_{\sigma
160: -})}] \nonumber \\
161: & & -c^{\sigma}_{12} \cos{(2\beta_{\rho}\theta_{\rho
162: -})}[2 \cos{(\beta_{\sigma}\Phi_{\sigma +})}+
163: \cos{(\beta_{\sigma}\Phi_{\sigma -})}] \nonumber \\
164: & & -c^{\sigma}_{12} \cos{(2\beta_{\rho}\theta_{\rho
165: -})}\cos{(\beta \theta_{\sigma -})} \nonumber \\
166: & & + 2 f^{\sigma}_{12} \cos{(\beta_{\sigma}\theta_{\sigma -})} \cos{(\beta_{\sigma}
167: \Phi_{\sigma +})}\: \} ,
168: \end{eqnarray}
169: where $\beta_{\rho}=\sqrt{\pi}$, $\beta_{\sigma}=-\sqrt{4\pi}$, and the fields $\Phi_{\nu\pm}=(\Phi_{\nu 1}\pm \Phi_{\nu
170: 2})/\sqrt{2}$ and $\Pi_{\nu\pm}=(\Pi_{\nu 1}\pm \Pi_{\nu 2})/\sqrt{2}$ have
171: been introduced. Upon minimizing the energy in a semiclassical approximation, any coupling that diverges
172: under the RG flow opens up a gap for a field that is pinned by the corresponding terms in (\ref{bosonhami}).
173:
174: % Furthermore, the
175: % quadratic part of the Hamiltonian, $H_q$, can be diagonalized as
176: %\begin{equation}
177: %\bar{H}_q= \sum_{\nu,i } \:\frac{u_{\nu i}}{2} \int \!\! dx\: \frac{1}{K_{\nu
178: % i}} (\partial_x
179: % \bar{\Phi}_{\nu i})^2 + K_{\nu i} \bar{\Pi}_{\nu i}^2,
180: %\end{equation}
181: %where
182: %$u_{\nu i}$ define the velocities of the low-energy excitations, and the
183: % Luttinger liquid parameters, $K_{\nu i}$, govern the long-distance
184: %behaviour of the various correlations functions.\cite{balents}
185: %Under the RG tranformations in the present
186: %the diagonalized fields, $\bar{\Phi}_{\mu i}$, are either
187: %rotated away from the original $\bar{\Phi}_{\mu i}$ with a mixing of the order of $U/t \ll 1$, or
188: %correspond to the (anti-) symmetric combinations $\Phi_{\mu \pm}$.
189:
190: Performing the above procedure, four different
191: phases are obtained for the Hamiltonian of Eq. (\ref{hubbardham}), shown in
192: the ($\delta$, $\Delta\mu)$-plane
193: for isotropic hopping, $t'=t$, in Fig. 5.
194: The
195: various phases are labeled according to the number of gapless charge (n) and spin (m) modes by
196: CnSm.
197: The different asymptotic regimes of the
198: RG flow (\ref{RGE}) are related to the phases shown in Fig. 5 as follows:
199: \begin{figure}
200: \includegraphics[width=\linewidth]{graphics/figRG1.eps}
201: \caption{Phase diagram of the inhomogeneously doped Hubbard
202: model on two-leg ladder in the weakly interacting limit, and for $t'=t$. The
203: phases CnSm are labeled according to
204: the number of gapless charge (n) and spin (m) modes. Solid lines are
205: phase boundaries. The dashed line indicates the crossover from a
206: region with dominant SC correlations in the lower and CDW and SDW
207: correlations in the upper band (C2S2.I), to a region dominated by CDW and SDW correlations
208: in both bands, for larger values of $\Delta\mu$ (C2S2.II).}
209: \label{figRG1}
210: \end{figure}
211:
212: (C1S1) The single band LL. This phase with an empty upper
213: band is labeled C1S1, reflecting the number of gapless modes. For incommensurate
214: filling the dominant
215: correlations are charge density waves (CDW) and spin density waves (SDW).\cite{schulz}
216:
217: (C2S2) The trivial fixed point.
218: In this regime the couplings stay
219: of order $U$ under the RG flow, or renormalize to
220: zero. Therefore no
221: gap opens in this phase, labeled C2S2.
222: Furthermore, the two partially filled bands are decoupled in this regime.
223: Standard LL theory\cite{schulz}, used to determine the dominant
224: correlations within each band, indicates a crossover between two regions labeled
225: C2S2.I, and C2S2.II respectively, c.f. Fig. 5. The dominant correlations in the C2S2.II regime are
226: CDW and SDW within both bands,
227: whereas in the C2S2.I regime the lower band is dominated by superconducting (SC)
228: fluctuations.
229:
230: (C2S1) Single-band superconductivity.
231: Here, all couplings stay of the order of $U$ or renormalize to zero, except for
232: $c^{\sigma}_{22} \sim -v_2$. This results in a pinning of the spin
233: mode of the lower band, and the number of gapless modes reduces to
234: C2S1. SC correlations dominate the lower band, and CDW and SDW
235: correlations the upper band. Furthermore, inter-band phase coherence is not
236: established within this regime.
237:
238: (C1S0) Inter-band superconductivity.
239: In this regime the diverging couplings flow towards the asymptotic ratios
240: \begin{equation}
241: 4c^{\rho}_{12}=8f^{\rho}_{12}=c^{\sigma}_{12}, \quad c_{11}^{\sigma}/v_1=c_{22}^{\sigma}/v_2.
242: \end{equation}
243: From the low-energy effective bosonic Hamiltonian two finite spin gaps are
244: obtained, and a
245: pinning of the charge mode
246: $\theta_{\rho -}=0$, resulting in
247: phase coherence between the two bands.
248: The number of gapless modes is reduced to C1S0.
249: The remaining total charge mode, ($\Phi_{\rho +}, \theta_{\rho +}$), is
250: gapless and exhibits dominant superconducting pairing correlations, with a
251: sign difference between the bands. This is usually referred to as the d-wave
252: -like superconducting phase of the two-leg ladder.\cite{balents,lin}
253:
254: The phase diagram in Fig. 5 confirms the results obtained along the line
255: $\Delta\mu=0$.\cite{balents,lin}
256: But it also indicates a limited stability of the various phases found at
257: $\Delta\mu=0$ under inhomogeneously doping of the ladder.
258: Upon increasing $\Delta\mu$, superconductivity is gradually suppressed,
259: with intermediate phases showing residual superconducting fluctuations.
260: Consider for example the low doping region where inter-band d-wave superconductivity
261: occurs for vanishing $\Delta\mu$. While for small
262: $\Delta\mu>0$ d-wave superconductivity sustains,
263: inter-band phase coherence is lost when $\Delta\mu$ reaches a value of
264: approximately $1.5 t'$. For larger values of $\Delta\mu$ intra-band
265: superconductivity persists within the
266: lower band (which predominately projects onto the lower
267: leg of the ladder). In the upper band spin fluctuations have become gapless
268: and SDW and CDW correlations dominate. Further increase of $\Delta\mu$ results
269: in the suppression of all superconducting
270: correlations, giving
271: rise to two-band LL behavior. In the weakly interacting limit
272: the upper band is depleted
273: for $\Delta\mu/t>3.5$, and a single-band LL,
274: residing mainly on the lower leg of the ladder, dominates the
275: large-$\Delta\mu$ regime.
276: This progressive reduction of superconducting pairing correlations is also observed
277: in the finite-temperature phase diagram. In Fig. 6 we show results for $t'=t$,
278: and $\Delta\mu/t=0.3$ in the $(\delta, T)$-plane. From the renormalization of the energy scale, $E \sim t e^{-\pi l}$, the logarithmic length scale of
279: Eq. (\ref{RGE}) can be related to a temperature scale $T=E=T_0 e^{-\pi l}$.
280: While the phases of the system for $T\rightarrow 0$ are found in accordance with
281: Fig. 5, the finite temperature phase
282: diagram reveals a successive enhancement of superconducting pairing
283: correlations with decreasing temperature.
284: Consider again the behavior close to half filling.
285: At high temperatures the system is dominated by LL behavior in both
286: bands. Upon decreasing the temperature
287: gapless superconducting correlations develop within the lower band. At even
288: lower temperatures, a finite
289: spin gap opens for the lower band, then finally phase coherent inter-band d-wave
290: superconductivity emerges, along with the opening of the second spin gap.
291: Thus in an intermediate temperature regime,
292: well above the onset of d-wave superconductivity, a single spin gap persists in the
293: lower band, which is related to the bonding band at small values of $\Delta\mu$. This partial spin gap
294: formation might be
295: interpreted as a phenomenon similar to the pseudogap phase in the HTCS
296: materials.
297:
298:
299: \begin{figure}[t]
300: \includegraphics[width=\linewidth]{graphics/figRG2.eps}
301: \caption{Finite Temperature phase diagram of the inhomogeneously doped Hubbard
302: model on the two-leg ladder in the weakly interacting limit for $t'=t$, and
303: $\Delta\mu/t=0.3$.
304: Solid lines are phase boundaries, whereas the dashed line indicates a
305: crossover inside the gapless regime. }
306: \label{figRG2}
307: \end{figure}
308:
309:
310:
311:
312:
313:
314:
315:
316:
317: