Incommensurate pair-density-wave correlations in two-leg ladder t t – J J – J ⟂ J_{\perp} model Hanbit Oh, Julian May-Mann, Ya-Hui Zhang Abstract We report the discovery of a generalized Luther-Emery liquid phase characterized by incommensurate pair-density-wave (iC-PDW) correlations in the two-leg t t - J J - J ⟂ J_{\perp} ladder model. By tuning the potential difference between the legs, we explore the regime of intermediate layer polarization P P . Combining density-matrix renormalization group (DMRG) simulations with bosonization analysis, we identify a spin-gapped phase at finite P P , where the interlayer and intralayer pair correlations both oscillate, but with distinct periodicities. The interlayer correlations exhibit FFLO-like oscillations, driven by pairing between layers with mismatched Fermi momenta, with a period determined by their momentum difference. In contrast, the intralayer pair correlations arise from the coupling between charges on one layer and spin fluctuations on the opposite layer, with a momentum equal to twice the Fermi momentum of the opposite layer. The iC-PDW state is robust across a wide range of doping and polarization, although finite interlayer hopping eventually destabilizes it toward a state with charge- 4 ​ e 4e correlations. We conclude by discussing the experimental realization of this model in optical lattice platforms and its relevance to the bilayer nickelate La 3 Ni 2 O 7 . Introduction.— Pair-density-wave (PDW) superconductivity, characterized by Cooper pairing at finite center-of-mass momentum, has been one of the central themes in the study of strongly correlated systems Agterberg et al. 2020 . A notable example of finite-momentum pairing is the Fulde–Ferrell–Larkin–Ovchinnikov (FFLO) phase, where pairing is driven by an external Zeeman field Fulde and Ferrell 1964 ; Larkin 1964 . On the other hand, a time-reversal-invariant PDW needs an intrinsically strong-coupling pairing mechanism to compete with the uniform superconductor. One well-established realization is in the one-dimensional Kondo–Heisenberg model, which exhibits a commensurate PDW at momentum q=\pi together with a finite spin gap Sikkema et al. 1997 ; Berg et al. 2010 . There is also a wide range of theoretical and numerical studies which have investigated PDW phases in various microscopic models Cho et al. 2014 ; May-Mann et al. 2020 ; Zhang and Vishwanath 2022 ; Nikolaenko and Zhang 2024 ; Liu et al. 2024 ; Huang et al. 2022 ; Jiang and Yao 2024 ; Chen et al. 2024 ; Liu and Han 2024 ; Chen and Sheng 2023 ; Jiang 2023 ; Peng et al. 2021 ; Venderley and Kim 2019a ; Xu et al. 2019 ; Venderley and Kim 2019b ; Soto-Garrido et al. 2015 ; Yang and Feiguin 2024 . Meanwhile, incommensurate PDW phases are less explored. To our knowledge, incommensurate PDW correlations have not yet been unambiguously demonstrated even in numerical studies. In this work, we propose a two-leg ladder t – J – J_{\perp} model as an ideal platform to search for incommensurate pair-density-wave (iC-PDW) correlations. The model consists of two t – J chains coupled solely through an interlayer spin-exchange interaction J_{\perp} , in the absence of interlayer charge hopping t_{\perp} [Fig. 1 (a)]. At zero layer-polarization, it hosts a uniform interlayer superconducting state with dominant singlet pairing correlations, corresponding to a Luther–Emery liquid in one dimension Luther and Emery 1974 . Beyond one dimension, the same model has been widely studied in the context of the bilayer nickelate La 3 Ni 2 O 7 Sun et al. 2023 , where strong Hund’s coupling between the d_{x^{2}-y^{2}} and d_{z^{2}} orbitals generates an effective interlayer spin exchange. In this setting, previous theoretical studies have revealed that the ground state is an interlayer s -wave superconductor Oh and Zhang 2023 ; Qu et al. 2024 ; Lu et al. 2023 ; Lu et al. 2024 ; Yang et al. 2024 ; Lange et al. 2024 ; Schlömer et al. 2024 ; Yang et al. 2025 ; Wang et al. 2025 ; Oh et al. 2025a ; Oh et al. 2025b . Most previous studies are limited to the mirror-symmetric regime of equal layer densities, and the effects of symmetry breaking have been relatively less explored Fan and Vishwanath 2025 ; Oh and Zhang 2025 . Figure 1: (a) Illustration of the two-leg ladder t – J – J_{\perp} model. Particle numbers on the two legs are independently conserved, and we set N_{1}\geq N_{2} . (b) Polarization-imbalance–driven phase transitions at t\sim J\sim J_{\perp} . For 0J_{c} , where the spin-gapped iC-PDW is realized. The main finite-DMRG results are summarized in Fig. 2 and Fig. 3 . Fig. 2 (a) shows the spin gap, \Delta_{s}=E(S_{z}=1)-E(S_{z}=0) . For sufficiently large J>J_{c}=0.78 , a finite spin gap persists across the entire polarization range 0\leq P\leq 1 . Although Fig. 2 (a) only shows the cases n=7/8 and n=38/47 , the spin gap is robust over different fillings and generally increases with increasing electron density, as shown in the Supplemental Material (SM). By contrast, in the small- J regime, J0.78 (spin-gapless region). A.3 C. More DMRG results for spin-gapped regime J>J_{c} We present additional DMRG results for the spin-gapped region J>J_{c} . In the main text, we only show the Fourier transform of the intralayer pairing correlations on leg 2, which is the less populated leg. Here we show that the intralayer pairing correlations on leg 1 do not exhibit a sharp finite-momentum peak, in contrast to those on leg 2, which display a clear peak at 2k_{F,1} . Figure S3: Fourier transform of the intralayer pairing correlations on layer 1. We use a filling of n=7/8 with parameters t=1 and J_{\perp}=J=2 . No sharp peak of \Delta_{\parallel,1} is observed for any value of P . The dashed line indicates q=2k_{F,2} . In Fig. S4 , we show the pair-correlation data for various dopings and polarizations, demonstrating that the interlayer PDW correlation is the dominant one. We further evaluate other correlation functions. In the broad parameter range, we found \displaystyle\Delta_{\perp},\Delta_{\parallel},P_{z}: \displaystyle\ \text{Gapless}, (S1) \displaystyle c_{t},c_{b},S_{z}: \displaystyle\ \text{Gapped}. (S2) as illustrated in Fig. S5 . Figure S4: Pair correlation function from fDMRG t=1 , J_{\perp}=2 , with \chi=2000 . (a,b) are at n=7/8 and (c,d) are at n=38/47 . Figure S5: Correlation functions of various operators obtained from fDMRG at filling n=7/8 with parameters t=1 , J_{\perp}=2 , and J=1 , using a bond dimension \chi=2000 . The main panel shows the data on a log–log scale, while the inset presents the same correlations on a linear–log scale. A.4 D. Effect of interlayer hopping t_{\perp} In the main text, we show that the iC-PDW phase stable under t_{\perp} term. In Fig S6 , we present entanglement entropy plot to extract the central charge. Figure S6: iDMRG results at filling n=7/8 with t=1 and J=J_{\perp}=2 . A layer-dependent potential \epsilon=0.6 is applied to induce a finite polarization \langle P\rangle\neq 0 . The simulations are performed with bond dimensions up to \chi\leq 5500 . (a) Entanglement entropy S as a function of correlation length \xi . (b) Central charge c extracted from the scaling relation c=6\,\partial S/\partial(\log\xi) as a function of \xi . For t_{\perp}=0.1 and 0.25 , we find c\simeq 2 over a broad range of correlation lengths. For larger t_{\perp} , however, the extracted central charge shows a tendency to deviate from c=2 and approach c=1 at larger bond dimensions or longer correlation lengths, suggesting a possibility to flow toward a c=1 phase at sufficiently large t_{\perp} . Appendix B II. Details of the Bosonization analysis Following the convention of Ref. Giamarchi 2003 , the electron operator is represented as \displaystyle c_{\ell,\sigma}(x)=R_{\ell,\sigma}(x)\,e^{ik_{F,\ell}x}+L_{\ell,\sigma}(x)\,e^{-ik_{F,\ell}x}, (S3) with \displaystyle r_{\ell,\sigma}(x) \displaystyle= \displaystyle\frac{1}{2\pi\alpha}\,U_{r,\ell,\sigma}\,\exp\!\Bigg[-\frac{i}{\sqrt{2}}\Big(r\phi_{\ell,c}-\theta_{\ell,c}+\sigma(r\phi_{\ell,s}-\theta_{\ell,s})\Big)\Bigg], (S4) where r=R,L labels right- and left-moving modes, \ell=1,2 is the layer index, and \sigma=\uparrow,\downarrow is the spin index. The fields \theta_{\ell,\rho} and \phi_{\ell,\rho} ( \rho=c,s ) are real bosonic fields describing charge and spin degrees of freedom, respectively: \displaystyle\phi_{\ell,c} \displaystyle= \displaystyle\frac{1}{\sqrt{2}}\big(\phi_{\ell,\uparrow}+\phi_{\ell,\downarrow}\big),\quad\phi_{\ell,s}=\frac{1}{\sqrt{2}}\big(\phi_{\ell,\uparrow}-\phi_{\ell,\downarrow}\big), (S5) \displaystyle\theta_{\ell,c} \displaystyle= \displaystyle\frac{1}{\sqrt{2}}\big(\theta_{\ell,\uparrow}+\theta_{\ell,\downarrow}\big),\quad\theta_{\ell,s}=\frac{1}{\sqrt{2}}\big(\theta_{\ell,\uparrow}-\theta_{\ell,\downarrow}\big), (S6) with the standard commutation relation \big[\phi_{\ell,\rho}(x),\partial_{y}\theta_{\ell^{\prime},\rho^{\prime}}(y)\big]=i\pi\,\delta_{\ell,\ell^{\prime}}\,\delta_{\rho,\rho^{\prime}}\,\delta(x-y). (S7) Here U_{r,\ell,\sigma} are Klein factors ensuring the correct anticommutation relations between fermions. At J_{\perp}=0 , the Hamiltonian density of the two-leg t – J – J_{\perp} model reduces to \displaystyle\mathcal{H}_{0} \displaystyle=\sum_{\ell=1,2}\left\{\frac{v_{\ell,c}}{2\pi}\Big[K_{\ell,c}(\partial_{x}\theta_{\ell,c})^{2}+\frac{1}{K_{\ell,c}}(\partial_{x}\phi_{\ell,c})^{2}\Big]+\frac{v_{\ell,s}}{2\pi}\Big[(\partial_{x}\theta_{\ell,s})^{2}+(\partial_{x}\phi_{\ell,s})^{2}\Big]\right\}. (S8) Here K_{\ell,c} is the charge Luttinger parameter determined by t , J , and the filling n_{\ell} , while v_{\ell,c} and v_{\ell,s} denote the charge and spin velocities, respectively. To express the interlayer spin interaction, we first bosonize the spin operators on each leg: \displaystyle S_{\ell}^{z}(x) \displaystyle= \displaystyle-\frac{1}{\sqrt{2}\pi}\partial_{x}\phi_{\ell,s}+\frac{1}{2\pi\alpha}\sin\!\big(\sqrt{2}\,\phi_{\ell,s}\big)\Big[e^{-2ik_{F,\ell}x}e^{i\sqrt{2}\phi_{\ell,c}}+e^{2ik_{F,\ell}x}e^{-i\sqrt{2}\phi_{\ell,c}}\Big], (S9) \displaystyle S_{\ell}^{+}(x) \displaystyle= \displaystyle\frac{e^{-i\sqrt{2}\theta_{\ell,s}}}{2\pi\alpha}\Big\{i\Big[e^{-2ik_{F,\ell}x}e^{i\sqrt{2}\phi_{\ell,c}}+e^{2ik_{F,\ell}x}e^{-i\sqrt{2}\phi_{\ell,c}}\Big]+2\sin\!\big(\sqrt{2}\,\phi_{\ell,s}\big)\Big\}, (S10) where we follow the choice of Klein factors in Ref. Zhang and Vishwanath 2022 . The interlayer spin interaction is then written as \displaystyle\mathcal{H}_{\perp} \displaystyle= \displaystyle g\Big[S^{z}_{1}(x)S^{z}_{2}(x)+\frac{1}{2}\Big(S^{+}_{1}(x)S^{-}_{2}(x)+S^{-}_{1}(x)S^{+}_{2}(x)\Big)\Big] (S11) \displaystyle\simeq \displaystyle g\left[\frac{1}{2\pi^{2}}\partial_{x}\phi_{1,s}\,\partial_{x}\phi_{2,s}+\frac{1}{4\pi^{2}\alpha^{2}}\sin\!\big(\sqrt{2}\,\phi_{1,s}\big)\sin\!\big(\sqrt{2}\,\phi_{2,s}\big)\cos\!\big(\sqrt{2}(\theta_{1,s}-\theta_{2,s})\big)\right] (S12) \displaystyle= \displaystyle g\left[\frac{1}{4\pi^{2}}\big((\partial_{x}\phi_{+,s})^{2}-(\partial_{x}\phi_{-,s})^{2}\big)-\frac{1}{2\pi^{2}\alpha^{2}}\big(\cos 2\phi_{+,s}-\cos 2\phi_{-,s}\big)\cos 2\theta_{-,s}\right], (S13) where we have kept only the slowly varying contributions and set g=J_{\perp} . The coupling g renormalizes the quadratic part of the Hamiltonian. In the following, we assume that the bare differences between the two legs are not substantial, so that K_{\ell,c}=K_{c} and v_{\ell,c}=v_{c} for both layers. The full Hamiltonian is then conveniently written in the \{\phi_{\pm,c/s},\,\theta_{\pm,c/s}\} basis as \displaystyle\mathcal{H} \displaystyle= \displaystyle\frac{v_{c}}{2\pi}\Big[K_{c}\left(\partial_{x}\theta_{+,c}\right)^{2}+\frac{1}{K_{c}}\left(\partial_{x}\phi_{+,c}\right)^{2}\Big]+\frac{v_{c}}{2\pi}\Big[K_{c}\left(\partial_{x}\theta_{-,c}\right)^{2}+\frac{1}{K_{c}}\left(\partial_{x}\phi_{-,c}\right)^{2}\Big] (S14) \displaystyle+\frac{v_{+,s}}{2\pi}\Big[K_{+,s}\left(\partial_{x}\theta_{+,s}\right)^{2}+\frac{1}{K_{+,s}}\left(\partial_{x}\phi_{+,s}\right)^{2}\Big]+\frac{v_{-,s}}{2\pi}\Big[K_{-,s}\left(\partial_{x}\theta_{-,s}\right)^{2}+\frac{1}{K_{-,s}}\left(\partial_{x}\phi_{-,s}\right)^{2}\Big] \displaystyle-\frac{g}{2\pi^{2}\alpha^{2}}\cos\!\left(2\phi_{+,s}\right)\cos\!\left(2\theta_{-,s}\right), where the renormalized spin-sector parameters are given by \displaystyle v_{\pm,s} \displaystyle= \displaystyle v_{s}\sqrt{1\pm\frac{g}{2\pi v_{s}}},\quad K_{\pm,s}=\frac{1}{\sqrt{1\pm\frac{g}{2\pi v_{s}}}}. (S15) In Eq. S14 we have discarded the term \cos 2\phi_{-,s}\,\cos 2\theta_{-,s} , whose coupling has scaling dimension [g^{\prime}]=2-K_{-,s}+\frac{1}{K_{-,s}}<0 and is therefore irrelevant in the RG sense. This vertex does not affect the low-energy physics and can be safely neglected. The one-loop RG equation of the remaining couplings are given by \displaystyle\frac{dK_{+,s}}{dl} \displaystyle= \displaystyle-\frac{1}{4}K_{+,s}^{2}\,g^{2}, (S16) \displaystyle\frac{dK_{-,s}}{dl} \displaystyle= \displaystyle\frac{1}{4}g^{2}, (S17) \displaystyle\frac{dg}{dl} \displaystyle= \displaystyle\left[2-K_{+,s}-\frac{1}{K_{-,s}}\right]g. (S18) For any initial value g(0)>0 , the coupling flows to g\to+\infty . At this strong-coupling fixed point, the spin fields \phi_{+,s} and \theta_{-,s} are pinned, opening a finite spin gap. Consequently, only the charge sector remains gapless, and the low-energy theory has central charge c=2 . B.1 A. Properties of order parameters We now analyze the long-distance behavior of various order parameters at the strong-coupling fixed point ( g\to\infty ). At this point, \phi_{+,s} and \theta_{-,s} are pinned, whereas their dual fields \theta_{+,s} and \phi_{-,s} are disordered. First, consider the interlayer PDW operator, \displaystyle\mathcal{O}_{\mathrm{PDW,\perp}} \displaystyle= \displaystyle c_{1,\uparrow}\,c_{2,\downarrow}-c_{1,\downarrow}\,c_{2,\uparrow}. (S19) The q=\delta k_{F} component of this operator can be decomposed as \displaystyle R_{1,\uparrow}L_{2,\downarrow} \displaystyle= \displaystyle\frac{1}{2\pi\alpha}\,e^{-i(\phi_{-,c}-\theta_{+,c}+\phi_{+,s}-\theta_{-,s})}=\frac{1}{2\pi\alpha}\,e^{i(-\phi_{-,c}+\theta_{+,c})}, (S20) \displaystyle R_{1,\downarrow}L_{2,\uparrow} \displaystyle= \displaystyle\frac{1}{2\pi\alpha}\,e^{-i(\phi_{-,c}-\theta_{+,c}-\phi_{+,s}+\theta_{-,s})}=\frac{1}{2\pi\alpha}\,e^{i(-\phi_{-,c}+\theta_{+,c})}, (S21) \displaystyle L_{1,\uparrow}R_{2,\downarrow} \displaystyle= \displaystyle\frac{1}{2\pi\alpha}\,e^{+i(\phi_{-,c}+\theta_{+,c}+\phi_{+,s}+\theta_{-,s})}=\frac{1}{2\pi\alpha}\,e^{i(\phi_{-,c}+\theta_{+,c})}, (S22) \displaystyle L_{1,\downarrow}R_{2,\uparrow} \displaystyle= \displaystyle\frac{1}{2\pi\alpha}\,e^{+i(\phi_{-,c}+\theta_{+,c}-\phi_{+,s}-\theta_{-,s})}=\frac{1}{2\pi\alpha}\,e^{i(\phi_{-,c}+\theta_{+,c})}. (S23) Hence the interlayer PDW operator is expressed purely in terms of the gapless charge fields, \displaystyle\mathcal{O}_{\mathrm{PDW,\perp}}\sim e^{i(\pm\phi_{-,c}+\theta_{+,c})}, (S24) and therefore exhibits quasi-long-range order. The corresponding two-point function decays as \displaystyle\big\langle\mathcal{O}^{\dagger}_{\mathrm{PDW,\perp}}(0)\,\mathcal{O}_{\mathrm{PDW,\perp}}(r)\big\rangle\sim|r|^{-\alpha_{\perp}}, (S25) with exponent \displaystyle\alpha_{\perp}=\frac{1}{2}\left[K_{-,c}+\frac{1}{K_{+,c}}\right]. (S26) We next consider the intralayer composite PDW order of layer 2, \displaystyle\mathcal{O}_{\mathrm{PDW},\parallel}^{(2)} \displaystyle\sim \displaystyle\vec{N}^{(1)}\cdot\vec{\Delta}^{(2)}_{T}. (S27) Here, \vec{\Delta}^{(2)}_{T} denotes the triplet pairing operator on leg 2, and \vec{N}^{(1)} is the 2k_{F,1} component of the spin operator on leg 1 [Eqs. S9 – S10 ], which continuously connects to the Néel order at P=1 . One finds \vec{\Delta}_{T}^{(2)}\sim e^{i\sqrt{2}\theta_{2,c}}(\sin\sqrt{2}\theta_{2,s},\cos\sqrt{2}\theta_{2,s},\sin\sqrt{2}\phi_{2,s}) and \displaystyle N_{+}^{(1)} \displaystyle= \displaystyle\frac{i}{2\pi\alpha}\,e^{-i\sqrt{2}\theta_{1,s}}e^{-i\sqrt{2}\phi_{1,c}}=N_{x}^{(1)}+iN_{y}^{(1)}, (S28) \displaystyle N_{-}^{(1)} \displaystyle= \displaystyle\frac{-i}{2\pi\alpha}\,e^{+i\sqrt{2}\theta_{1,s}}e^{-i\sqrt{2}\phi_{1,c}}=N_{x}^{(1)}-iN_{y}^{(1)}, (S29) \displaystyle N_{z}^{(1)} \displaystyle= \displaystyle\frac{1}{2\pi\alpha}\sin\sqrt{2}\phi_{1,s}\,e^{-i\sqrt{2}\phi_{1,c}}, (S30) so that \vec{N}^{(1)}\sim e^{-i\sqrt{2}\phi_{1,c}}(\sin\sqrt{2}\theta_{1,s},\cos\sqrt{2}\theta_{1,s},\sin\sqrt{2}\phi_{1,s}) . The composite order at layer 2 then bosonizes as \mathcal{O}_{\mathrm{PDW},\parallel}^{(2)}\sim e^{i\sqrt{2}\theta_{2,c}}\,e^{-i\sqrt{2}\phi_{1,c}}\sim e^{i(\theta_{+,c}-\theta_{-,c}-\phi_{+,c}-\phi_{-,c})}, (S31) which again involves only the gapless charge fields and thus exhibits quasi-long-range correlations. Similarly, we find that the bosonized form of \displaystyle\mathcal{O}_{\mathrm{PDW},\parallel}^{(1)} \displaystyle\sim \displaystyle\vec{N}^{(2)}\cdot\vec{\Delta}^{(1)}_{T} (S32) is given by \mathcal{O}_{\mathrm{PDW},\parallel}^{(1)}\sim e^{i\sqrt{2}\theta_{1,c}}\,e^{-i\sqrt{2}\phi_{2,c}}\sim e^{i(\theta_{+,c}+\theta_{-,c}-\phi_{+,c}+\phi_{-,c})}. (S33) The bosonized form of \mathcal{O}_{\mathrm{PDW},\parallel}^{(1)} differs from that of \mathcal{O}_{\mathrm{PDW},\parallel}^{(2)} only by a sign factor in the exponential. The corresponding two-point function decays as \displaystyle\big\langle\mathcal{O}^{\dagger}_{\mathrm{PDW,\parallel}}(0)\,\mathcal{O}_{\mathrm{PDW,\parallel}}(r)\big\rangle\sim|r|^{-\alpha_{\parallel}},\quad\ell=1,2, (S34) with exponent \displaystyle\alpha_{\parallel}=\frac{1}{2}\left[K_{+,c}+K_{-,c}+\frac{1}{K_{+,c}}+\frac{1}{K_{-,c}}\right]. (S35) From the above analysis, the scaling dimensions of the intralayer PDW orders in the two layers are found to be the same. This conclusion is based on the simple assumption that the Luttinger parameters of each layer are the same, K_{1}=K_{2} . In general, however, it is very likely that K_{1}\neq K_{2} in realistic situations, and the scaling dimensions of the two intralayer components can therefore be different, as demonstrated in our DMRG results. B.2 B. Effect of interlayer hopping t_{\perp} We now briefly comment on the effect of a small interlayer hopping in the UV theory H_{t_{\perp}}=-t_{\perp}\sum_{i,\sigma}\left(c^{\dagger}_{i1\sigma}c_{i2\sigma}+\text{H.c.}\right), (S36) which leads to the continuum Hamiltonian \displaystyle\mathcal{H}_{t_{\perp}} \displaystyle\sim \displaystyle-t_{\perp}\sum_{\sigma}\Big[R^{\dagger}_{1\sigma}R_{2\sigma}\,e^{-i\delta k_{F}x}+L^{\dagger}_{1\sigma}L_{2\sigma}\,e^{i\delta k_{F}x} (S37) \displaystyle\qquad\qquad+\,R^{\dagger}_{1\sigma}L_{2\sigma}\,e^{-i(k_{F,1}+k_{F,2})x}+L^{\dagger}_{1\sigma}R_{2\sigma}\,e^{i(k_{F,1}+k_{F,2})x}+\text{H.c.}\Big]. Due to the mismatch between the Fermi momenta of the two legs, this single-particle tunneling rapidly oscillate. Consequently, these terms average to zero under coarse-graining and are effectively RG-irrelevant. We further consider higher-order processes that emerge as the theory flows from the UV to the IR. Specifically, at second order in t_{\perp} , a two-particle pair-hopping process is generated: \displaystyle\mathcal{H}_{2p} \displaystyle\sim \displaystyle g_{2p}\int dx\left(c_{1\uparrow}^{\dagger}c_{1\downarrow}^{\dagger}c_{2\downarrow}c_{2\uparrow}+\text{H.c.}\right), (S38) where the coupling constant scales as g_{2p}\sim t_{\perp}^{2} . Unlike single-particle tunneling, this pair-hopping operator remains commensurate. This operator can be expressed as: \displaystyle c_{1\uparrow}^{\dagger}c_{1\downarrow}^{\dagger}c_{2\downarrow}c_{2\uparrow}+\text{H.c.} \displaystyle\sim \displaystyle\cos(\sqrt{2}\theta_{-,c}-\sqrt{2}\phi_{+,s}). (S39) Note that the spin-sector involves only the \phi_{+,s} , which is pinned as 0 or \pi in the iC-PDW phase. Thus the pair-hopping term is purely determined by the relative charge field \theta_{-,c} . The scaling dimension of the pair-hopping coupling g_{2p} is given by: \displaystyle[g_{2p}] \displaystyle= \displaystyle 2-\frac{1}{K_{-,c}}. (S40) For K_{-,c}>1/2 , the pair-hopping term becomes relevant and additionally pins \theta_{-,c} field. Consequently, the corresponding mode is gapped out, remaining a single gapless mode, resulting in a c=1 phase. Pinning of \theta_{-,c} gaps out charge 2e PDW correlations, as they involve the dual field, \phi_{-,c} which highly fluctuates (see Eqs. S24 , S31 ). Instead, the dominant gapless mode is now a charge- 4e superconducting order, defined as \Delta_{4e}\sim\mathcal{O}_{\mathrm{PDW},\perp}^{2} , where \displaystyle\mathcal{O}_{\mathrm{PDW},\perp}\sim e^{i(k_{F,1}-k_{F,2})x}\,e^{-i\phi_{-,c}+i\theta_{+,c}}+e^{-i(k_{F,1}-k_{F,2})x}\,e^{i\phi_{-,c}+i\theta_{+,c}}. (S41) Squaring this operator, one finds a uniform contribution that is independent of the relative charge field \phi_{-,c} , \displaystyle\Delta_{4e}\sim e^{2i\theta_{+,c}}. (S42) As a result, while interlayer 2e PDW correlations are suppressed, the uniform charge- 4e correlations remain long-ranged decaying as \displaystyle\langle\Delta_{4e}^{\dagger}(0)\Delta_{4e}(r)\rangle\sim|r|^{-\frac{2}{K_{+,c}}}. (S43)