betaFeedback

WorksPhysics and Astronomy

Strain-tunable van Hove singularity and superconducting dome in monolayer kagome metal CsTi3_3Bi5_5

Abstract

The recently discovered titanium-based kagome metal CsTi3_3Bi5_5, which exhibits superconductivity without an accompanying charge density wave, provides an ideal platform to investigate the intrinsic electronic structure and superconductivity. Moreover, it enables the exploration of superconductivity in the two-dimensional limit of the kagome lattice. Here, we systematically investigate the influence of biaxial mechanical strain (ε\varepsilon) on the electronic structure and superconductivity of monolayer CsTi3_3Bi5_5 using first-principles calculations. We demonstrate that the prominent van Hove singularity (VHS) at the M point can be monotonically tuned to the Fermi level by tensile strain. Furthermore, the superconducting transition temperature TCT_C exhibits a dome-shaped dependence on strain, reaching a maximum of 3.6 K at an optimal tensile strain of ε=+2%\varepsilon=+2\%. The average superconducting gap magnitude is 0.35 meV, with an orbital-dependent and anisotropic distribution. Our results provide a detailed understanding of the superconductivity of monolayer CsTi3_3Bi5_5 and show that strain is an effective, continuous tuning knob for manipulating the electronic structure and superconductivity.

Introduction

The kagome lattice is well known for its geometric frustration and unique electronic structure. The recently discovered kagome superconductor AV₃Sb₅ (A=K, Rb, Cs) exhibits exotic quantum orders, including charge density wave (CDW) [1, 2, 3], superconductivity [4, 5, 6], nematicity [7, 8], pair density wave [9, 10], and nontrivial topology [11]. Dimensionality acts as an independent tuning parameter in kagome metals, significantly affecting their charge order and superconductivity. In exfoliated CsV₃Sb₅, for thicknesses below about 10 layers, the CDW transition temperature increases while the superconducting transition temperature TCT_C decreases, and the material becomes insulating below 5 layers [12]. The newly discovered titanium-based kagome metal CsTi₃Bi₅, which exhibits superconductivity without CDW order, offers a clean platform to study intrinsic kagome superconductivity in the two-dimensional limit [13, 14].

CsTi₃Bi₅ exhibits a distinct electronic structure compared with its vanadium-based counterparts. A notable feature is the van Hove singularity (VHS) just above the Fermi level. In contrast to CsV₃Sb₅, the VHS in CsTi₃Bi₅ can be tuned to the Fermi level by chemical doping [15]. However, chemical doping inevitably introduces disorder, so a clean and continuous tuning knob is needed to probe the intrinsic kagome superconductivity. Lattice strain, typically originating from substrate mismatch or mechanical bending, offers such an approach. Strain engineering has indeed proven effective at modulating the electronic structure and superconductivity of diverse two-dimensional materials [16, 17, 18, 19]. Recently, strain experiments on CsV₃Sb₅ have shown that its nodal pairing can be enhanced by applying strain [20]. Whether strain can effectively tune the electronic structure and superconductivity in the CDW-free kagome metal CsTi₃Bi₅ remains unexplored.

In this work, we systematically investigate the influence of strain on the electronic structure and superconductivity of monolayer CsTi₃Bi₅ using first-principles calculations. Our results show that strain can be used as a clean and continuous knob to modulate the electronic structure, Fermi surface topology, and superconductivity of monolayer CsTi₃Bi₅. We find that as the in-plane lattice parameter expands, the apical Bi atoms approach the kagome plane, similar to the effect of applying pressure in the bulk. The VHS located at the M point can be monotonically tuned to the Fermi level by tensile strain, while compressive strain shifts it to higher energy. A dome-shaped dependence of TCT_C on strain is observed, with TCT_C reaching a maximum of 3.6 K at an optimal tensile strain of +2%. The calculated superconducting gap exhibits a strongly anisotropic and orbital-dependent distribution over the Fermi surfaces, with an average magnitude of 0.35 meV at ε=+2%\varepsilon= +2\%. The planar Bi-6pz6p_z orbital yields smaller gap values (∼\sim0.32 meV), while the Ti-3d3z2,xz,yz3d_{3z^2,xz,yz} and 3dxy,x2−y23d_{xy,x^2-y^2} orbitals give larger gap values (0.35-0.42 meV). Our results not only provide a detailed understanding of the superconductivity of monolayer CsTi₃Bi₅ but also show that strain is an effective, continuous tuning knob for manipulating electronic structure and superconductivity.

Methods

The first-principles calculations based on density functional theory were performed using the Quantum ESPRESSO package [21, 22]. Ultrasoft pseudopotentials were used to describe the interaction between electrons and ionic cores [23]. The exchange-correlation interaction was described within the generalized gradient approximation using the Perdew-Burke-Ernzerhof functional [24]. The cutoffs of kinetic energy and charge density were set to 50 and 500 Ry, respectively. The van der Waals (vdW) interaction was included via Grimme’s DFT-D3 method [25]. A vacuum layer of 20 Å thickness was introduced to construct the monolayer slab model. The structure was relaxed until the Hellmann-Feynman force acting on each atom was less than 10−410^{-4} Ry/Å, and the convergence criterion for self-consistent calculations was set to 10−810^{-8} Ry. The Brillouin zone was sampled with 20×20×420 \times20 \times4 and 5×5×15 \times5 \times1 kk- and qq-point grids for electronic and phonon calculations, respectively. The superconducting gaps were calculated using the EPW code [26]. The electron-phonon matrix was interpolated onto 40×40×440 \times40 \times4 kk- and qq-point grids. The Matsubara frequency cutoff was set to 1 eV, and the Dirac delta functions were replaced by Lorentzians with widths of 100 and 0.5 meV for electrons and phonons, respectively. The Coulomb pseudopotential was set to μ∗=0.1\mu^{*} = 0.1 in both the Allen-Dynes and anisotropic Migdal-Eliashberg calculations.

Results and Discussion

The titanium-based kagome metal CsTi3_3Bi5_5 exhibits a layered structure in the P6/mmmP6/mmm space group, identical to that of the vanadium-based kagome superconductors of the AV3_3Sb5_5 family [13]. Figure 1(a) shows the structure of monolayer CsTi3_3Bi5_5, where the kagome lattice is formed by Ti and planar Bi atoms. The Bi atoms in the kagome plane and those outside it are denoted as planar and apical Bi, respectively. The band structure is also similar to that of CsV3_3Sb5_5, but its Fermi energy (EFE_{\mathrm{F}}) is lower (see Fig. S1). We begin by investigating the unstrained structure to provide a baseline for understanding strain effects on the electronic structure and superconductivity of monolayer CsTi3_3Bi5_5. Figure 1(c) shows the orbital-resolved band structure. Four bands cross the Fermi level, originating from the Ti-3dxz,yz3d_{xz,yz}, Ti-3dxy,x2−y23d_{xy,x^{2}-y^{2}}, apical Bi-6px,y6p_{x,y} and planar Bi-6pz6p_{z} orbitals. The α\alpha band close to the Γ\Gamma point originates from the Bi-6pz6p_{z} orbital, forming a cylindrical Fermi surface (FS). The hexagonal β\beta and γ\gamma FSs are primarily composed of the Ti-3dxz,yz3d_{xz,yz} and 3dxy,x2−y23d_{xy,x^{2}-y^{2}} orbitals, respectively, with a minor 6px,y6p_{x,y} contribution from apical Bi to both sheets (see Fig. S3(d)). The δ\delta FS close to the M and K points is mainly derived from Ti-3dxz,yz3d_{xz,yz} orbitals.

Structure, Fermi surface, orbital-resolved band structure, phonon spectra, and projected phonon spectra of monolayer CsTi3Bi5

Figure 1. (a) Structure of monolayer CsTi3_3Bi5_5. (b) Fermi surface. (c) Orbital-resolved band structure. (d) Phonon spectra and integrated electron-phonon coupling (EPC) strength λ\lambda. The width of the blue lines indicates the mode-resolved EPC strength λqν\lambda_{q\nu}. (e) Phonon spectra projected onto the vibrational modes of Ti and apical Bi atoms.

In the unstrained structure, the calculated total EPC strength λ\lambda is 0.54, as shown in the right panel of Figure 1(d). The width of the blue lines on the phonon spectra indi-

cates the mode-resolved EPC strength (λqν\lambda_{q\nu}), calculated by

λqν=2ℏNFωqν×∑j∫BZdkΩBZ∣gqν(k,j,j)∣2δ(ϵk,i−EF)δ(ϵk+q,j−EF)(1)\lambda_{q\nu} = \frac{2}{\hbar N_{\mathrm{F}}\omega_{q\nu}} \times\sum_{j} \int_{BZ} \frac{d k}{\Omega_{BZ}} \lvert g_{q\nu}(k,j,j) \rvert^{2} \delta(\epsilon_{k,i} - E_{\mathrm{F}}) \delta(\epsilon_{k+q,j} - E_{\mathrm{F}}) \tag*{(1)}

where gqνg_{q\nu} stands for the EPC matrix elements. The phonon modes with large λqν\lambda_{q\nu} originate primarily from the low-frequency out-of-plane vibrations of the apical Bi atoms and the high-frequency in-plane vibrations of the Ti atoms (see Fig. S4 for all atomic contributions). Based on the Allen-Dynes modified McMillan equation

Tc=ωlog⁡1.2exp⁡[−1.04(1+λ)λ−μ∗−0.62λμ∗](2)T_{\mathrm{c}} = \frac{\omega_{\log}}{1.2} \exp\left[-\frac{1.04(1+\lambda)}{\lambda-\mu^{*}-0.62\lambda\mu^{*}}\right] \tag*{(2)}

the estimated superconducting transition temperature (TCT_{\mathrm{C}}) is 2.6 K.

To investigate the influence of strain on the electronic structure and superconductivity, we apply biaxial strain in the abab plane by simultaneously adjusting the lattice parameters aa and bb. The strain magnitude is defined as ϵ=(a−a0)/a0×100%\epsilon= (a-a_{0})/a_{0} \times100\%, where a0a_{0} is the unstrained lattice constant. Positive (negative) values of ϵ\epsilon thus denote tensile (compressive) strain. The in-plane lattice parameters are fixed according to the applied strain ϵ\epsilon, while the out-of-plane zz-coordinates of the apical Bi and Cs atoms are fully relaxed to minimize the total energy. Since the Cs atoms move along zz in the same way as the neighboring apical Bi atoms and contribute negligibly to the total EPC strength, we plot only the distance dd between the two apical Bi atoms in Figure 2(b). Increasing the strain ϵ\epsilon monotonically drives the apical Bi atoms toward the kagome plane, an effect analogous to that of hydrostatic pressure on the bulk crystal. In addition, this biaxial distortion enlarges the lattice parameter aa, thus expanding the kagome plane and increasing the distance between Ti atoms. Mechanical strain therefore produces effects that are partly analogous to, and partly distinct from, those of hydrostatic pressure.

Lateral and top views of monolayer CsTi3Bi5, changes in apical Bi distance, and band structures under strain

Figure 2. (a) Lateral and top views of monolayer CsTi3_3Bi5_5. (b) Changes in the distance between apical Bi atoms. (c)-(e) Evolution of the band structure with applied strain ε\varepsilon. The black arrow highlights the VHS located at the M point.

We calculate the band structures under various strain conditions, ranging from compressive (ε=−4%\varepsilon= -4\%) to tensile (ε=+6%\varepsilon= +6\%). The unstrained band structure (ε=0%\varepsilon= 0\%) shows the VHS above EFE_F at the M point, composed of the Ti-3dxy,x2−y23d_{xy,x^2-y^2} orbitals (blue lines). Under compressive strain (Figure 2(c)), the VHS shifts to higher energy, moving further from the Fermi level. The hole-like band around the M point, which arises from apical Bi-6px,y6p_{x,y} and Ti-3dxz,yz3d_{xz,yz} orbitals, shifts to lower energy. Conversely, tensile strain (Figure 2(d) and (Figure 2e)) causes the VHS to shift toward lower energy. At ε=+6%\varepsilon= +6\%, the VHS is tuned precisely to the Fermi level. The flat band below EFE_F likewise approaches the Fermi level. This monotonic, strain-driven shift of the VHS at the M point to the Fermi level (highlighted by black arrows) demonstrates that the electronic structure of this kagome metal can be manipulated by mechanical strain. Such a strain-tunable VHS offers a route to probing its relationship with competing electronic orders.

To elucidate the influence of strain on the superconductivity of monolayer CsTi3_3Bi5_5, we calculate the phonon spectra and the associated electron-phonon coupling for various strain values. Figures 3(a)-(c) present the calculated phonon dispersions, showing that the phonon modes soften as the strain increases from compressive to tensile values. The mode- and momentum-resolved EPC strength (λqν\lambda_{\mathbf{q}\nu}) is represented by the width of the blue lines. As strain increases from −4%-4\% to +2%+2\%, the mode-resolved EPC strength λqν\lambda_{\mathbf{q}\nu} of the two modes marked by black arrows increases (Figure 3(a)). These two modes are dominated by the out-of-plane vibrations of the apical Bi atoms (Figure 3(f)) and the in-plane vibrations of the Ti atoms (Figure 3(g)), respectively. Integration of the Eliashberg spectral function α2F(ω)\alpha^2F(\omega) gives the total EPC

Phonon spectra, electron-phonon coupling, and phonon modes under applied strain

Figure 3. (a)-(c) Evolution of the phonon spectra with applied strain ε\varepsilon. The width of blue lines indicates EPC strength λqν\lambda_{\mathbf{q}\nu}. (d) Total EPC strength λ\lambda and superconducting transition temperature TCT_C versus strain ε\varepsilon. (e) Frequency-integrated EPC strength λ(ω)\lambda(\omega). (f)-(g) Vibrational schematics of corresponding phonon modes.

strength λ\lambda, which increases from 0.37 at ε=−4%\varepsilon= -4\% to 0.62 at ε=+2%\varepsilon= +2\% (Figure 3(e)). Beyond +2%+2\%, λ\lambda decreases to 0.46 at ε=+6%\varepsilon= +6\%. The nonmonotonic strain dependence of EPC, which first increases and then decreases, gives rise to a dome-shaped dependence of TCT_C on strain. As summarized in Figure 3(d), TCT_C, estimated using the Allen-Dynes modified McMillan equation, rises from 0.3 K at −4%-4\% to a maximum of 3.6 K at +2%+2\%, and then falls to 1.2 K at +6%+6\%.

To gain a deeper insight into the superconductivity in monolayer CsTi3_3Bi5_5, we calculate the momentum-resolved superconducting gap Δnk\Delta_{\mathbf{n}\mathbf{k}} using the anisotropic Migdal-Eliashberg theory. Figure 4(a) shows the distribution of the superconducting gap on the Fermi surface under the optimal tensile strain (ε=+2%\varepsilon= +2\%). The gap is fully open, with an average magnitude of 0.35 meV. The central pocket is dominated by the planar Bi-6pz6p_z orbital, exhibiting the smallest gap magnitude of 0.32 meV. The hexagonal Fermi surfaces β\beta and γ\gamma are dominated by Ti-3dxz,yz3d_{xz,yz} and 3dx2−y23d_{x^2-y^2} orbitals, respectively, with gap values ranging from 0.35 to 0.42 meV. The sheet of FS around the K point exhibits moderate gap magnitudes and also mainly arises from Ti-3dxz,yz3d_{xz,yz} orbitals. Solving the anisotropic Migdal-Eliashberg equations at different temperatures gives the temperature dependence of Δnk\Delta_{\mathbf{n}\mathbf{k}}, from which we estimate TC=3.2T_C = 3.2 K, close to the TCT_C estimated by the Allen-Dynes modified McMillan equation.

Momentum-resolved superconducting gap and distributions around the Fermi surface

Figure 4. (a) Momentum-resolved superconducting gap Δnk\Delta_{n\mathbf{k}} on the Fermi surface, calculated by anisotropic Migdal-Eliashberg theory. (b) Energy distribution of the superconducting gap Δnk\Delta_{n\mathbf{k}} as a function of temperature. (c), (d) Distribution of EPC strength λnk\lambda_{n\mathbf{k}} and superconducting gap Δnk\Delta_{n\mathbf{k}} around the Fermi surface.

Kagome metals such as the AV3_3Sb5_5 family have attracted intense interest in their superconductivity. The recently discovered CsTi3_3Bi5_5 has been reported to exhibit superconductivity without structural instability, thereby providing an ideal system to access the intrinsic superconductivity of the kagome lattice. Our first-principles calculations systematically demonstrate that biaxial strain is an effective and clean tool to manipulate the electronic structure and significantly enhance the electron-phonon-mediated superconductivity in monolayer CsTi₃Bi₅. The VHS at the M point can be continuously tuned to the Fermi energy by tensile strain, culminating in a precise alignment at ε=+6%\varepsilon= +6\%. Furthermore, TCT_C can be enhanced to 3.6 K at the optimal strain ε=+2%\varepsilon= +2\%. This value is higher than the ambient-pressure bulk TCT_C of 1.85 K estimated within the first-principles framework [14]. Notably, the VHS reaches the Fermi level at ε=+6%\varepsilon= +6\%, whereas TCT_C is maximized at ε=+2%\varepsilon= +2\%. This indicates that the TCT_C enhancement is not

determined solely by the VHS-EFE_F alignment but also by the strain dependence of λ\lambda. Finally, the resulting orbital dependence of the gap resembles the behavior inferred for CsV₃Sb₅, suggesting a common origin across titanium- and vanadium-based kagome superconductors.

Conclusion

In conclusion, through first-principles calculations, we have systematically investigated the influence of biaxial strain on the electronic structure and superconductivity of monolayer CsTi₃Bi₅. Notably, the VHS at the M point can be monotonically tuned to the Fermi level by applying tensile strain. In addition, tensile strain can significantly enhance TCT_C from 0.3 K to 3.6 K as the strain varies from -4% to +2%, and then falls to 1.2 K at +6%, giving a dome-shaped dependence of TCT_C on strain. Within the anisotropic Migdal-Eliashberg framework, the calculated superconducting gap attains an average magnitude of 0.35 meV at ε=+2%\varepsilon= +2\% and exhibits an orbital-dependent and anisotropic distribution. More generally, these results position biaxial strain as a disorder-free and continuously tunable knob for exploring how lattice geometry controls pairing in two-dimensional kagome metals.

Acknowledgments

This work was supported by the National Key Research and Development Program of China under Contract No. 2022YFA1403203.

References

  1. [1]L. Zheng, Z. Wu, Y. Yang, L. Nie, M. Shan, K. Sun, D. Song, F. Yu, J. Li, D. Zhao, S. Li, B. Kang, Y. Zhou, K. Liu, Z. Xiang, J. Ying, Z. Wang, T. Wu, and X. Chen, Nature 611, 682 (2022).
  2. [2]Z. W. Liang, X. Y. Hou, F. Zhang, W. R. Ma, P. Wu, Z. Y. Zhang, F. H. Yu, J. J. Ying, K. Jiang, L. Shan, Z. Y. Wang, and X. H. Chen, Phys. Rev. X 11, 031026 (2021).
  3. [3]B. R. Ortiz, S. M. L. Teicher, L. Kautzsch, P. M. Sarte, N. Ratcliff, J. Harter, S. Seshadri, and S. D. Wilson, Phys. Rev. X 11, 041030 (2021).
  4. [4]K. Jiang, T. Wu, J. X. Yin, Z. Wang, M. Z. Hasan, S. D. Wilson, X. Chen, and J. Hu, Natl. Sci. Rev. 10, nwac199 (2023).
  5. [5]K. Y. Chen, N. N. Wang, Q. W. Yin, Y. H. Gu, K. Jiang, Z. J. Tu, C. S. Gong, Y. Uwatoko, J. P. Sun, H. C. Lei, J. P. Hu, and J. G. Cheng, Phys. Rev. Lett. 126, 247001 (2021).
  6. [6]H. S. Xu, Y. J. Yan, R. Yin, W. Xia, S. Fang, Z. Chen, Y. Li, W. Yang, Y. Guo, and D. L. Feng, Phys. Rev. Lett. 127, 187004 (2021).
  7. [7]Y. Sur, K. T. Kim, S. Kim, and K. H. Kim, Nat. Commun. 14, 3899 (2023).
  8. [8]L. Nie, K. Sun, W. Ma, D. Song, L. Zheng, Z. Liang, P. Wu, F. Yu, J. Li, M. Shan, D. Zhao, S. Li, B. Kang, Z. Wu, Y. Zhou, K. Liu, Z. Xiang, J. Ying, Z. Wang, T. Wu, and X. Chen, Nature 604, 59 (2022).
  9. [9]H. Chen, H. Yang, B. Hu, Z. Zhao, J. Yuan, Y. Xing, G. Qian, Z. Huang, G. Li, Y. Ye, S. Ma, S. Ni, H. Zhang, Q. Yin, C. Gong, Z. Tu, H. Lei, H. Tan, S. Zhou, C. S. Chen, X. Dong, B. Yan, Z. Wang, and H. J. Gao, Nature 599, 222 (2021).
  10. [10]J. T. Jin, K. Jiang, H. Yao, and Y. Zhou, Phys. Rev. Lett. 129, 167001 (2022).
  11. [11]B. R. Ortiz, S. M. L. Teicher, Y. Hu, J. L. Zuo, P. M. Sarte, B. R. Ortiz, C. Schueller, A. M. Abeykoon, M. J. Krogstad, S. Rosenkranz, R. Osborn, R. Seshadri, L. Balents, J. He, and S. D. Wilson, Phys. Rev. Lett. 125, 247002 (2020).
  12. [12]B. Song, T. Ying, X. Wu, W. Xia, Q. Yin, Q. Zhang, Y. Song, X. Yang, J. Guo, L. Gu, X. Chen, J. Hu, A. P. Schnyder, H. Lei, Y. Guo, and S. Li, Nat. Commun. 14, 2492 (2023).
  13. [13]The vanadium-based kagome superconductors of the AV3Sb5 family.
  14. [14]X. W. Yi, Z. W. Liao, J. Y. You, B. Gu, and G. Su, Research 6, 0238 (2023).
  15. [15]B. Liu, M. Q. Kuang, Y. Luo, Y. Li, C. Hu, J. Liu, Q. Xiao, X. Zheng, L. Huai, S. Peng, Z. Wei, J. Shen, B. Wang, Y. Miao, X. Sun, Z. Ou, S. Cui, Z. Sun, M. Hashimoto, D. Lu, C. Jozwiak, A. Bostwick, E. Rotenberg, L. Moreschini, A. Lanzara, Y. Wang, Y. Peng, Y. Yao, Z. Wang, and J. He, Phys. Rev. Lett. 131, 026701 (2023).
  16. [16]Y. P. Qi, M. A. Sadi, D. Hu, M. Zheng, Z. P. Wu, Y. C. Jiang, and Y. P. Chen, Adv. Mater. 35, 2205714 (2023).
  17. [17]Z. Liu, Z. Zhang, H. Song, T. Cui, and D. Duan, Chin. Phys. B 34, 047104 (2025).
  18. [18]T. F. Poon, K. Y. Yip, Y. K. Tsui, L. Wang, K. H. Yu, W. Zhang, Z. Wang, T. Nakatani, C. Michioka, H. Ueda, S. T. Lam, K. T. Lai, and S. K. Goh, APL Mater. 13, 061121 (2025).
  19. [19]C. Lin, A. Consiglio, O. K. Forslund, J. Küspert, M. M. Denner, H. Lei, A. Louat, M. D. Watson, T. K. Kim, C. Cacho, D. Carbone, M. Leandersson, C. Polley, T. Balasubramanian, D. D. Sante, R. Thomale, Z. Guguchia, G. Sangiovanni, T. Neupert, and J. Chang, Nat. Commun. 15, 10466 (2024).
  20. [20]Y. Takeuchi, A. Kobayashi, S. Uchida, T. Nagao, S. Ogawa, R. Zhou, S. Kawasaki, F. Song, H. Ni, Y. Zhao, and G. Q. Zheng, Phys. Rev. Lett. 137, 096003 (2026).

Paper details

vlnc:W_95pn4nvk
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