$d_{x^{2}-y^{2}}$-wave Bose Metal induced by the next-nearest-neighbor hop** $t^{\prime}$

Abstract

Superconductivity arises when electrons form Cooper pairs with phase coherence. In contrast, a lack of phase coherence in Cooper pairs can lead to an uncondensed metallic ground state known as the Bose metal state. In this study, we investigate an attractively interacting fermionic system with nearest-neighbor (NN) hop** (t) and next-nearest-neighbor (NNN) hop** ($t^{\prime}$) anisotropy between two species of spins in a two-dimensional (2D) lattice. Utilizing the constrained path quantum Monte Carlo (CPQMC) method, we demonstrate the existence of a $d_{x^{2}-y^{2}}$-wave Cooper pair Bose metal (CPBM) phase with $t^{\prime}/t>0.7$. The CPBM phase exhibits a dome-like structure in the phase diagram of filling $n\sim 0.65$, with the maximal region around an optimal $t^{\prime}/t\sim 0.2$, suggesting that an appropriate value of $t^{\prime}$ facilitates the formation of the Bose metal. Furthermore, we find that a Bose metal formed by fermions with a closed Fermi surface confirms that the crucial condition for this exotic phenomenon is primarily the anisotropy of the Fermi surface, rather than its topology. Our finding of the $d_{x^{2}-y^{2}}$-wave CPBM demonstrates the same pairing symmetry as the pseudogap behavior in cuprates, and its experimental realization in ultracold atom systems is also feasible.

The cuprate superconductors [1] have sparked significant interest over the past three decades not only due to their high-temperature ($T_{c}$) unconventional superconductivity. The peculiar behavior of the pseudogap (PG) phenomena in high-$T_{c}$ superconductors, which might be closely related to the microscopic mechanism of superconductivity, still awaits a well-recognized explanation [2, 3, 4]. One of the most puzzling fact is that the onset of PG phase is accompanied with a gap with “$d_{x^{2}-y^{2}}$-wave like” symmetry opening below the characteristic temperature ($T^{*}$) with no superconductivity signal  [5].

Various types of order are proposed to explain the $d$-wave-like PG behavior. In the famous resonating valence bond (RVB) theory [6, 7], the quantum spin liquid (QSL) is argued as the origin of the PG phase, where the charge and spin degrees of freedom are separated, defines a PG of spinons below $T^{*}$ when holon condensation is absent [8]. As an alternation of the QSL phase, the Bose liquid or Bose metal is also propsed to explain the strange metal behavior in PG phase, which presumes that Cooper pairs are dominant charge carriers for the electric transport not only in the superconducting (SC) but also in the metallic phases, constituting a conducting quantum fluid instead of a superfluid [9].

In other words, the existence of a Bose metal phase, which Cooper pairs instead of electrons as the primary charge carriers is a bosonic system [10], could potentially explain the strange metal state in high-$T_{c}$ superconductors. Indeed, from Bardeen-Cooper-Schrieffer (BCS) to Bose-Einstein condensation (BEC) crossover theory [11], incoherent Cooper pairs begin to form at temperatures $T_{c}<T<T^{*}$, before they Bose condense and superconduct at $T_{c}$, which is at the heart of BCS-BEC crossover theory.

Bose metal is argued to leave fingerprint in microscopic model of hard-core bosons with ring exchange on multileg ladders, which called bosonic $J-K$ model [12, 13]. As a variant of Bose metals, Cooper pair Bose metal (CPBM) has been theoretically proposed to exist in 2D systems with no polarization [14, 15]. The anisotropic spin-dependent Fermi surface plus attractive interactions lead to an effective model of Cooper pairs with a ring-exchange term, that may allow to realize a paired but non-superfluid Bose metal phase. The Cooper pairs would form a collective state with gapless excitations along a Bose surface but no condensate in momentum space.

Notably, the fingerprint of CPBM phase in fermions system has been observed in the qausi-1D systems like two-leg ladder [15] and four-leg ladder [16]. In our recent work [17], we have found the CPBM phase in the 2D $t-U$ Hubbard model, with onsite attraction interaction ($U$) and the nearest-neighbor (NN) spin-dependent hop** ($t$) anisotropy, in which fermions are paired as bosons, and the uncondensed Cooper pairs form a non-superfluid Bose metal phase. But the boson correlation of Bose metal is mainly the $d_{xy}$-wave, induced by $t$. In a previous study [18], it was suggested that adding next-nearest-neighbor (NNN) hop** ($t^{\prime}$) anisotropy might induce the $d_{x^{2}-y^{2}}$-wave correlation in the Bose metal phase.

The NNN hop** $t^{\prime}$ is also argued to strongly affect the properties of superconductivity and other orders, as has been revealed in the previous studies in the Hubbard model [19, 20, 21, 22, 23, 24, 25]. A delicate interplay between superconductivity and density wave orders tunable via NNN hop** $t^{\prime}$ is revealed by using extensive density matrix renormalization group (DMRG) studies of the $t-t^{\prime}-U$ Hubbard model at hole do** concentration $\delta=0.125$ on four-leg cylinders [22]. The semi-classical Monte Carlo calculations show that introduction of $t^{\prime}$ results in a finite temperature PG phase that separates the small $U$ Fermi liquid and large $U$ Mott insulator. We have discovered the presence of Bose metal phase in $t-U$ Hubbard model [17], but the influence of the NNN hop** $t^{\prime}$ on the ground state properties of Bose metal phase remains an outstanding theoretical question.

Here, we choose to focus on the $t-t^{\prime}-U$ Hubbard model, utilizing the constrained path quantum Monte Carlo (CPQMC) [26, 27], to look at the effect of $t^{\prime}$ on the Bose metal phase. Our key conclusions are: (i) We found the exotic CPBM phase in $t-t^{\prime}-U$ Hubbard model, emerging when a spin-dependent anisotropy suppresses the ordinary $s$-wave pairing, and emphasize that the appropriate value of $t^{\prime}$ facilitates the formation of the Bose metal phase; (ii) our analysis identifies primary competition between $d_{xy}$-wave and two types of $d_{x^{2}-y^{2}}$-wave within the CPBM phase, when $t^{\prime}$ is large enough ($t^{\prime}/t>0.7$), the system becomes dominated by $d_{x^{2}-y^{2}}$-wave Bose metal phase; (iii) we find that a Bose metal formed by fermions with a closed Fermi surface confirms that the crucial condition for this exotic phenomenon is primarily the anisotropy of the Fermi surface, rather than its topology. This $d_{x^{2}-y^{2}}$-wave Bose metal phase with uncondensed bosons in 2D, provides a controllable route to enhance CPBM and control the closeness of the Fermi surface, offering insights into the theoretical understanding of the pseudogap phase.

Figure 1: (Color online) Model and phase diagram. (a) Illustration of the model with spin-dependent anisotropic hop** amplitudes on a square lattice. (b) Schematic zero-temperature phase diagram at $U=-3$ (the unit is $t$) for electron filling $n\sim 0.65$, exotic Cooper pair Bose metal (CPBM) exists in the regime with strong spin-dependent anisotropy that suppresses the ordinary $s$-wave superfluid ($s$-SF). IDW is an incommensurate density wave that exists in strong anisotropy and small $t^{\prime}$. CPBM phase divided into two types with $t^{\prime}\sim 0.7$ as the boundary, and dominated by $d_{xy}$-wave and $d_{x^{2}-y^{2}}$-wave correlation between bosons, respectively. The inset shows the illustration for conventional $s$-wave pairing by up and down spins, and nonzero momentum $s$-wave pairing by up and down spins. (c)-(e) CPQMC simulation result of the $s$-wave pair momentum distribution function $N_{\mathrm{s}-pair}({\bf k})$ for different $t^{\prime}=0.2,\ 0.5,\ 0.8$ at $\alpha=0.05$ with electron filling $n\sim 0.65$ on 16 $\times$ 16 lattice, showing obvious existence of nonzero momentum Bose surface, which is the signal for determining the exotic CPBM phase. The determination of each phase region in the phase diagram is inferred from our CPQMC data (details provided in Supplemental Material [28]).

The Hamiltonian of $t-t^{\prime}-U$ Hubbard model on a square lattice is given by

$$H=-\sum_{\begin{subarray}{c}i,l,\sigma\end{subarray}}\left(t_{l,\sigma}c_{i,\sigma}^{\dagger}c_{i+l,\sigma}+h.c.\right)+U\sum_{i}n_{i,\uparrow}n_{i,\downarrow}$$ (1)

where $c_{i,\sigma}^{\dagger}$ ($c_{i,\sigma}^{\,}$) is electron creation (annihilation) operators with spin $\sigma$ = $\uparrow,\downarrow$, and $n_{i,\sigma}=c_{i,\sigma}^{\dagger}c_{i,\sigma}$ is the electron number operator. The electron hop** amplitude $t_{l,\sigma}$, where $l=\pm\hat{x},\pm\hat{y}$ represents the NN sites of a given site $i$, and $l=\pm\hat{x}\pm\hat{y}$ represents the NNN sites of a given site $i$. $U<0$ is the on-site attractive interaction. For defining the spin-dependent anisotropic hop** amplitudes, we define the variable $\alpha$ and $\alpha$’, where $t_{\hat{y}\downarrow}=t_{\hat{x}\uparrow}=t$, $t_{\hat{x}\downarrow}=t_{\hat{y}\uparrow}=\alpha t$ (we set the NN hop** $t$ = 1 as the energy unit) and $t_{\hat{x}-\hat{y}\downarrow}=t_{\hat{x}+\hat{y}\uparrow}=t^{\prime}$, $t_{\hat{x}+\hat{y}\downarrow}=t_{\hat{x}-\hat{y}\uparrow}=\alpha^{\prime}t^{\prime}$, leading to an unpolarized system with balanced spin populations $\langle n_{i,\uparrow}\rangle=\langle n_{i,\downarrow}\rangle=n/2$. We give a schematic diagram with anisotropic NN and NNN hop** on 2D square lattice (see Fig. 1a), without loss of generality we defined the anisotropy parameter $\alpha,\alpha^{\prime}\in[0,1]$. Such a Hubbard model with unequal hop** amplitudes can be readily implemented in an optical lattice by loading mixtures of ultracold fermionic atoms with different masses. The correlation functions of CPQMC are discussed in the Supplemental Material [28].

CPBM is an exotic non-superfluid paired state of fermions, the uncondensed Cooper pairs would form a collective state with low-energy gapless excitations along a Bose surface [14, 15]. In our recent work [17], we explore a diverse phase diagram as a function of electron density $n$ and anisotropy $\alpha$ in $t-U$ model, and reveal the emergence of a CPBM phase in a highly anisotropic regime with wide range of filling. These experiences are also applied to the $t-t^{\prime}-U$ model, so we now focus on a case with electron density $n\sim 0.65$ at $U=-3$, and explore the influence of NNN hop** $t^{\prime}$ on the ground state properties of the Bose metal phase. The CPBM phase in this parameter range is relatively stable on different lattice sizes. In this paper, we assume $\alpha=\alpha^{\prime}$ for simplicity.

As we demonstrate in Fig. 1(b), we observe a diverse phase diagram for different value of NNN hop** $t^{\prime}$ and anisotropy $\alpha$ with electron filling $n\sim 0.65$ at zero temperature. Overall, the exotic CPBM phase still exists in a strong anisotropy, emerging when a spin-dependent anisotropy suppresses the ordinary $s$-wave superfluid ($s$-SF). Fig. 1(c)-(e) shows the Bose surface in momentum space representing different values of $t^{\prime}=0.2,\ 0.5,\ 0.8$ at extreme anisotropy $\alpha=0.05$. IDW is an incommensurate density wave at extremely strong anisotropy range, and it disappears when $t^{\prime}>0.4$. We can see that the Bose surface still exists, and the appearance of $t^{\prime}$ does not significantly change the main properties of the CPBM phase, but it changed the shape and size of the Bose surface, which is mainly related to the Fermi surface distortion caused by spin anisotropy. However, the region of CPBM phase is affected by $t^{\prime}$, we can see from the phase diagram that as $t^{\prime}$ increases, the area where CPBM exists presents a dome-like shape, the maximal area of CPBM phase occurs at around the optimal strength of $t^{\prime}\sim 0.2$, and then decreases in both small and large $t^{\prime}$ regions. This indicates that the presence of the appropriate value of $t^{\prime}$ is beneficial for Bose metal phase. Further details on determining the various phase regions in the phase diagram from CPQMC data are presented in the Supplemental Material [28].

Figure 2: (Color online) $d$-wave boson correlation. (a) and (b) (left panel) The schematic of the two-boson (on site pairing) $d_{xy}$-wave and $d^{(2)}_{x^{2}-y^{2}}$-wave pairing in square lattice. (right panel) Two-boson correlator $P_{d_{xy}-boson}({\bf k})$, $P_{d^{(2)}_{x^{2}-y^{2}}-boson}({\bf k})$ for $t^{\prime}=0.2$ and $0.8$ at $\alpha=0.05$ with $n\sim 0.65$ on 16 $\times$ 16 lattice, respectively. (c) Processes that would give origin to an effective ring-exchange $K_{t}$ of pairs induced by $t$ anisotropy. (left) Fermions with opposite spin prefer to move along perpendicular directions. (d) Processes that would give origin to an effective ring-exchange $K_{t^{\prime}}$ of pairs induced by $t^{\prime}$ anisotropy. (left) Fermions with opposite spin prefer to move along diagonal directions perpendicularly.

Figure 3: (Color online) Competition between $d_{x^{2}-y^{2}}$-wave and $d_{xy}$-wave boson correlation. The strength of four pairing modes of bosons paired correlation at $\alpha=0.05$ with $n\sim 0.65$ varies with different $t^{\prime}$. The inset shows boson pairing correlations in the real space for $t^{\prime}=0.2$ and $0.8$, marked by the black dashed lines.

One of the most important prospect of the Bose metal phase is its $d$-wave correlation between bosons (on-site pairing) [15, 17]. We defined the two-boson correlator $P_{\mathrm{\zeta}-boson}({\bf k})$ to study the possibility of a phase of Cooper pairs. Here $\zeta=s,\ d_{xy},\ d^{(1)}_{x^{2}-y^{2}},\ d^{(2)}_{x^{2}-y^{2}}$. Among them, the $d_{x^{2}-y^{2}}$-wave is divided into two types, namely the NN and the third NN $d_{x^{2}-y^{2}}$-wave, which we use $d^{(1)}_{x^{2}-y^{2}}$-wave and $d^{(2)}_{x^{2}-y^{2}}$-wave to represent. Our previous results [17] suggest that the correlation between Cooper pairs is predominantly $d_{xy}$-wave in the CPBM regime when there is no $t^{\prime}$ in 2D system.

Figure 4: (Color online) Characteristic of CPBM phase. (a)-(c) Non-interacting energy spectrum of the spin-dependent Fermi surface with anisotropy $\alpha=0.05$ for $t^{\prime}=0.2,\ 0.5,\ 0.8$ at filling $n\sim 0.65$. The Fermi surface is projected as red (blue) lines for the up-spin (down-spin). We provide an illustration of pairing with nonzero momentum. (d) The $s$-wave pairing correlation function in the CPBM phase and the $s$-SF phase in real space. The solid blue line is an exponential fitting curve. The inset is lattice size effect in CPBM phase and $s$-SF phase.

In left panel of Fig. 2(a) and Fig. 2(b), we give the schematic of the two-boson $d_{xy}$-wave pairing correlations on the diagonals in square lattice and $d^{(2)}_{x^{2}-y^{2}}$-wave pairing correlations on the third NN sites. The $d$-wave here specifically refers to $d_{xy}$-orbital or $d_{x^{2}-y^{2}}$-orbital symmetry, which exhibits a propensity to introduce $d$-wave correlations into the system and qualitatively alter the sign structure of the electronic ground state. We show the CPQMC simulation results of two-boson correlator $P_{\mathrm{\zeta}-boson}({\bf k})$ in momentum space for $\alpha=0.05$ at different $t^{\prime}=0.2$ and $t^{\prime}=0.8$, $\zeta=\ d_{xy},\ d^{(2)}_{x^{2}-y^{2}}$, respectively. In the highly anisotropic limit, the process depicted in Fig. 2(c) and (d) dominates, corresponding to $d_{xy}$-wave and $d^{(2)}_{x^{2}-y^{2}}$-wave boson correlations induced by $t$ and $t^{\prime}$, respectively.

By model map**, we provided the connection between our model of fermions with spin-dependent anisotropic Fermi surface to the bosonic $J-K$ model. When $\left|U\right|\gg t$, all of the fermions are considered to be tightly bound into on-site Cooper pairs. We can derive an effective boson Hamiltonian by considering a perturbation expansion in powers of $t/\left|U\right|$ [29, 14, 17, 15]. For $t-t^{\prime}-U$ model, the effective Hamiltonian can be written as: $H_{b}=-J\sum_{i,j}b_{i}^{\dagger}b_{j}+K_{t}\sum_{ring}b_{1}^{\dagger}b_{2}b_{3}^{\dagger}b_{4}+K_{t^{\prime}}\sum^{\prime}_{ring}b_{1}^{\dagger}b_{2}b_{3}^{\dagger}b_{4}+h.c.$. Here, we consider the boson ring-exchange term $K_{t}$ and $K_{t^{\prime}}$ as depicted in Fig. 2(c) and (d), which involves two bosons on opposite corners of an elementary square plaquette and rhomboid plaquette rotating by $\pm 90$ degrees. Here, $b_{i}^{\dagger}=c_{i\uparrow}^{\dagger}c_{i\downarrow}^{\dagger}$, $K_{t}$ corresponds to $i=1,2,3,4$ labeling sites taken clockwise around a square plaquette, $K_{t^{\prime}}$ corresponds to $i=1,2,3,4$ labeling sites taken clockwise around a rhomboid plaquette. Remarkably, in the extreme anisotropic limit, the ring terms $K_{t}$ and $K_{t^{\prime}}$ which hops pairs of bosons is nonzero, they can even be comparable to the spin exchange coupling term $J$. At the same time as the $K_{t^{\prime}}$ ring exchange term induced by $t^{\prime}$ anisotropy, the boson in the ring exchange may also have a boson correlation with another NN boson. In addition, we also provide the physical mechanism for enhancing $d^{(1)}_{x^{2}-y^{2}}$-wave by $K_{t^{\prime}}$ in Supplemental Material [28]. Particularly, with increasing $t^{\prime}$, the coupling strength $K_{t^{\prime}}$ becomes increasingly more important in the system, leading to the $d^{(1)}_{x^{2}-y^{2}}$-wave and $d^{(2)}_{x^{2}-y^{2}}$-wave all becoming more important.

We demonstrated changes in the strength of four pairing modes of boson paired correlation by fixing $\alpha=0.05$ at $n\sim 0.65$ and changing the NNN hop** $t^{\prime}$ (Fig. 3). Overall, the primary competition arises between $d_{xy}$-wave and two types of $d_{x^{2}-y^{2}}$-wave, while $s$-wave holds relatively less significance with a small value and the three types of $d$-waves have relatively high values in CPBM region. As $t^{\prime}$ increases, the intensity of the $d^{(1)}_{x^{2}-y^{2}}$-wave boson correlation increases and the $d^{(2)}_{x^{2}-y^{2}}$-wave does not change much, the proportion of the $d_{xy}$-wave boson correlation decreases, and ultimately the $d_{x^{2}-y^{2}}$-wave dominates when $t^{\prime}>0.7$. In the inset of Fig. 3, we compare the correlations between four pairing modes of boson pairs in real space for $t^{\prime}=0.2$ and $0.8$ (the parameters consistent with right panel of Fig. 2(a) and Fig. 2(b)). They all decay rapidly to zero, indicating that the boson correlation is short-range. The strength of the $d_{xy}$-wave boson correlation is the strongest when $t^{\prime}=0.2$, while two types of $d_{x^{2}-y^{2}}$-wave dominate when $t^{\prime}=0.8$. Overall, we suggest its predominant boson pairing mode is always $d$-wave character in Bose metal phase. Particularly, our CPQMC results indicate that the system tends toward $d_{x^{2}-y^{2}}$-wave correlation between Cooper pairs when $t^{\prime}$ is large ($t^{\prime}>0.7$), and $t^{\prime}$ being a key tuning parameter to adjust the intensity of boson correlation in the boson metallic phase. This exotic phase of boson-paired states can be termed the $d$-wave Bose Metal [30, 15, 18, 17].

We show the non-interacting energy dispersion with anisotropy $\alpha=0.05$ for $t^{\prime}=0.2,\ 0.5,\ 0.8$ in Fig. 4(a)-(c). The solid red and blue lines below indicate the anisotropic Fermi surface of spin-up and spin-down for a projection at $n\sim 0.65$. Due to the severe mismatch Fermi surface, fermions with different spin near the Fermi surface favor forming pairs at finite momentum. The previous mean-field study restricts the pairing between patches of fermions with opposite spin located along a specific direction [29]. Here our quantum Monte Carlo result considers the influence of all possible pairing channel with the restriction. We provide an illustration of nonzero momentum pairing in Fig. 4(a)-(c), respectively. Pairs with different finite momentum are formed by the spin-up and spin-down fermions at different parts of anistropic Fermi surface, constituting a continuous Bose surface with singular pair distribution function, as shown in Fig. 1(c)-(e), the maximum weight of pairing is relatively evenly distributed on the Bose surface. The shape and sharpness of the Bose surface are regulated by both the filling $n$ and hop** anisotropy $\alpha$ and the NNN hop** $t^{\prime}$.

The presence of anisotropy $\alpha$ induces a mismatch between the spin-up and spin-down Fermi surfaces, while $t^{\prime}$ distorts the shape of the Fermi surface. As $t^{\prime}$ increases, this distortion intensifies. When the distortion is significant, the rotation of the anisotropic Fermi surface from the $t^{\prime}=0$ limit becomes pronounced. At the $t^{\prime}=t$ limit, the elliptic Fermi surfaces almost orient along the $k_{x}\pm k_{y}$ direction. Consequently, the resulting Bose surface in the CPBM phase also rotates accordingly, as illustrated. This leads to closed Fermi surfaces forming at small to medium filling. Previous studies have pointed out that the openness of the Fermi surfaces is crucial in classifying the Bose metal, resulting in $d$-wave Local Bose Liquid (DLBL) with an open Fermi surface and $d$-wave Bose Liquid (DBL) with a closed Fermi surface [29]. The analogous CPBM phase to DBL can be realized by introducing a sufficiently large $t^{\prime}$, whereas previous studies of CPBM mostly correspond to DLBL due to the open nature of the Fermi surfaces.

Fig. 4(d) compares the CPQMC simulation results of the pair correlation function for CPBM phase and $s$-SF phase in real space. In particularly, the correlation in real space decay to zero quickly and fluctuate around zero with increasing distance in CPBM phase, showing that correlation is short-range, and there is no significant difference as $t^{\prime}$ changes. On the contrary, the correlation exhibits an exponential decay, and converges to a steady finite value at long distance in $s$-SF phase, this is the hallmark of the long-range superconducting correlation among Cooper pairs. Inset of Fig. 4(d) show the lattice size effect in different phases, we can see that in CPBM region when $\alpha=0.05$, the value of $N_{\mathrm{s}-pair}({\bf k}_{\rm max})$ has always been stable, which represents that the CPBM phase will always exist in large or even infinite lattice size. When $\alpha=1.00$, the system is $s$-SF phase, the value of $N_{\mathrm{s}-pair}({\bf k}_{\rm max})$ is continuously increasing, so $s$-SF is significantly diverge at weak anisotropic region.

In recent years, charge order or charge density wave (CDW)—a static periodic modulation of charge density and lattice positions driven by the Fermi surface, has been shown to be a universal property of cuprates [31] and other unconventional superconductors [32]. IDW along the lattice diagonals exists at strong anisotropy, displaying singular features condenses at nonzero momentum point $\bf Q$ = ($2k_{F}$, $2k_{F}$), gradually diffusing to point $\bf Q$ = ($\pi$, $\pi$) convert to CDW as $n$ approached to half filling [17]. We also explore the influence of $t^{\prime}$ on density wave and associated periodicity. We demonstrate that IDW disappears when $t^{\prime}>0.4$ in phase diagram, suggesting that too large $t^{\prime}$ is not conducive to the formation of IDW. We have discussed the influence of $t^{\prime}$ on density wave and associated periodicity in details in the Supplemental Material [28].

To conclude, we utilize the CPQMC algorithm to study the effect of the NNN hop** $t^{\prime}$ on the CPBM phase and its boson correlation in 2D lattice. In the highly anisotropic regime ($\alpha<0.40$) at filling n$\sim$0.65 for various $t^{\prime}$ values, we observe the presence of Bose surface in Cooper-pair distribution function, which is compelling evidence of the CPBM phase. Furthermore, the CPBM region in phase diagram exhibits a dome-like shape, with the maximal CPBM phase region occurring at around the optimal strength of $t^{\prime}\sim 0.2$, and then decreases in both small and large $t^{\prime}$ regions. Subsequently, we explored the existence of boson correlation, suggesting that the boson correlation between Cooper pairs is predominantly $d_{x^{2}-y^{2}}$-wave in the CPBM regime with large $t^{\prime}$ ($t^{\prime}>0.7$). We also argued that the necessary condition to form CPBM phase is the anisotropy of the Fermi surface caused by strong spin anisotropy.

Recently, spin-dependent anisotropic Fermi surfaces have garnered significant attention in condensed matter physics, particularly in the context of altermagnetism [33, 34]. A recent study theoretically proposed how a $d$-wave altermagnetic phase can be realized with ultracold fermionic atoms in optical lattices, which in a altermagnetic Hubbard model with anisotropic NNN hop** $t^{\prime}$ [35]. The $d_{x^{2}-y^{2}}$-wave Bose metal phase can be realized in optical lattice experiments by tuning the filling and hop** anisotropy of effective spin interactions with light. Our results demonstrate that the properties of Bose metal phase is strongly affected by modifications of the NNN hop** $t^{\prime}$ in Hubbard model. This provides valuable guidance for ultracold atomic gases in optical lattices, which can be microscopically engineered and measured to investigate these exotic phases within specific parameter ranges.

This work is supported by the National Natural Science Foundation of China (Grant No. 12204130), Shenzhen Start-Up Research Funds (Grant No. HA11409065), HITSZ Start-Up Funds (Grant No. X2022000), Shenzhen Key Laboratory of Advanced Functional CarbonMaterials Research and Comprehensive Application (Grant No. ZDSYS20220527171407017). T.Y. acknowledges supports from Natural Science Foundation of Heilongjiang Province (No. YQ2023A004).