Shift current conductivity in monolayer SnS: a tight-binding analysis Yuki Kusunoki,1 Tomoaki Kameda,1 and Katsunori Wakabayashi1, 2, 3 1 arXiv:2606.02851v1 [cond-mat.mtrl-sci] 1 Jun 2026 Department of Nanotechnology for Sustainable Energy, School of Science and Technology, Kwansei Gakuin University, Gakuen-Uegahara 1, Sanda, 669-1330, Japan 2 Research Center for Materials Nanoarchitectonics (MANA), National Institute for Materials Science (NIMS), Namiki 1-1, Tsukuba 305-0044, Japan 3 Center for Spintronics Research Network (CSRN), Osaka University, Toyonaka 560-8531, Japan (Dated: June 3, 2026) We investigate the bulk photovoltaic effect in monolayer SnS using an effective tight-binding model derived from first-principles calculations. By comparing short-range and long-range hopping models, we show that the essential features of the shift current conductivity are captured by a minimal model. The shift current is decomposed into transition intensity and shift vector, enabling identification of dominant interband transitions. The comparison reveals that long-range hopping processes quantitatively modify the peak positions and magnitudes, while the short-range model retains the characteristic low-energy structure of the nonlinear response. Our findings provide a transparent framework for understanding and designing bulk photovoltaic effects in two-dimensional materials. I. ELECTRONIC STRUCTURE OF MONOLAYER SNS A. Crystal structure Monolayer tin sulfide (SnS) is a two-dimensional semiconductor belonging to the group-IV monochalcogenides (XS, XSe; X = Sn, Ge), which have attracted considerable interest due to their intrinsic ferroelectricity and strong in-plane anisotropy [12, 14]. In contrast to conventional ferroelectrics, monolayer SnS exhibits robust in-plane polarization originating from its puckered lattice structure along the armchair direction, making it particularly suitable for nanoscale device applications. The crystal structure of monolayer SnS is shown in Fig. 1. The unit cell consists of two Sn atoms and two S atoms arranged in a rectangular lattice. The puckered configuration breaks inversion symmetry and leads to a finite spontaneous polarization within the plane. Due to this puckered structure, monolayer SnS retains a large spontaneous polarization even at elevated temperatures, making it a promising candidate for ferroelectric and optoelectronic applications [12, 19]. B. Effective tight-binding model The electronic structure of monolayer SnS was investigated within the framework of density functional theory (DFT) using the Vienna Ab initio Simulation Package (VASP) [20, 21]. The projector augmented-wave (PAW) method [22] was employed, and the exchange–correlation functional was treated within the generalized gradient approximation (GGA) in the Perdew–Burke–Ernzerhof (PBE) form [23]. To avoid spurious interactions between periodic images, a vacuum region larger than 15 Å was introduced along the out-of-plane direction. The planewave cutoff energy was set to 440 eV, and the Brillouin zone (BZ) was sampled using a Γ-centered Monkhorst– Pack grid of 16×16×1 [24]. The total energy convergence criterion was set to 10−10 eV. To analyze the low-energy electronic structure relevant to optical transitions, we construct an effective tight-binding model based on maximally localized Wannier functions (MLWFs) [25–27]. The Wannier functions were generated using the Wannier90 code [28], starting σij (ω) = from the DFT band structure. The resulting Wannierinterpolated Hamiltonian provides an accurate description of the bands near the band gap while significantly reducing the computational cost for evaluating optical responses. Figure 1(d) shows the density of states (DOS) and the comparison between the DFT band structure and the Wannier-based tight-binding model. The excellent agreement demonstrates that the constructed model faithfully reproduces the essential features of the electronic structure in the energy window of interest. The constructed tight-binding model accurately captures the band gap and symmetry-protected degeneracies along the Y → S → X path in the Brillouin zone. The effective Hamiltonian is represented by a 12 × 12 matrix, reflecting the p-orbital manifold of Sn and S atoms that dominates the low-energy electronic structure. This Wannier-based formulation enables an efficient and physically transparent analysis of optical and nonlinear response functions in monolayer SnS. The details of the effective tight-binding models with different hopping ranges are described in Appendix A. III. LINEAR OPTICAL RESPONSE OF MONOLAYER SNS In this section, we investigate the linear optical response of monolayer SnS using the effective tight-binding models introduced in Sec. II B. To balance computational efficiency and accuracy, we consider two models with different hopping ranges: a short-range (SR) model including only nearest-neighbor hopping terms, and a long-range (LR) model incorporating extended hopping processes, as illustrated in Fig. 2. Figure 3 shows the corresponding band structures obtained from the two models. The LR model reproduces the dispersion near the band edges more accurately, while the SR model captures the essential low-energy features relevant to optical transitions. When an external electromagnetic field is applied, the induced current density is described within linear response theory as X (1) σij (ω)Ej (ω), (1) Ji (ω) = j where Ej (ω) is the electric field component. The linear optical conductivity σij (ω) is calculated using the Kubo formula [2, 29]: i~e2 X X f (Enk ) − f (Emk ) hunk |v̂i |umk ihumk |v̂j |unk i · , S Emk − Enk Emk − Enk − ~ω − iη n,m k where |unk i is the cell-periodic Bloch function, and f (Enk ) is the Fermi–Dirac distribution. (2) 3 (a) (b) top view side view (c) (d) DFT Wannier90 Sn-s Sn-p Sn-d S-s S-p S-d total DOS Sn S FIG. 1. Crystal structure and Brillouin zone of monolayer SnS. (a) Top view of the lattice, where the green rectangle denotes the primitive unit cell containing two Sn atoms and two S atoms. (b) Side views along the armchair and zigzag directions, highlighting the puckered geometry that breaks inversion symmetry and gives rise to in-plane ferroelectric polarization. (c) First Brillouin zone with high-symmetry points Γ, X, Y , and S used in the band structure calculations. (d) Energy band structure (left panel) and DOS (right panel) of monolayer SnS along high-symmetry points in the first Brillouin zone. Red circles represent the DFT results, while black lines denote the Wannier-based tight-binding model. (a) Sn (b) S The velocity operator is given by a2 a2 a1 D (3) B In the static limit, the transverse conductivity is related to the Berry curvature [30]: y x 1 ∇k Ĥ(k). ~ A C y v̂k = a1 x FIG. 2. Tight-binding models with different hopping ranges. (a) Short-range (SR) model including only nearest-neighbor hopping. (b) Long-range (LR) model including extended hopping processes. The green regions indicate the unit cells, and a1 and a2 denote the lattice vectors. σxy = e2 1 X z Ω (k), ~ S (4) k where the Berry curvature is defined as Ωz (k) = −2~2 Im X f (Enk ) n Although the Berry curvature is locally finite due to inversion symmetry breaking, its integral over the Brillouin zone vanishes, resulting in σxy = 0. This is consistent with time-reversal symmetry, which forbids an intrinsic Hall response in monolayer SnS [30]. Figure 4 summarizes the Berry curvature and the linear optical response of monolayer SnS. Figure 4(a) shows the Berry curvature Ωz (k) calculated using the LR model. The Berry curvature exhibits an anisotropic distribution in the first Brillouin zone, reflecting the reduced symmetry of monolayer SnS. Figure 4(b) compares the X hunk |v̂x |umk ihumk |v̂y |unk i m6=n (Emk − Enk )2 . (5) longitudinal optical conductivity σxx (ω) obtained from the SR and LR models. The SR model reproduces the main low-energy spectral features of the LR model, while the inclusion of longer-range hoppings modifies the peak positions and detailed spectral shapes. In particular, the peak heights in the linear optical conductivity are relatively similar between the two models compared with the shift-current spectra discussed in Sec. IV.B. The linear response formulation provides the basis for analyzing the nonlinear optical response, including the bulk photovoltaic effect, which is discussed in the follow- 4 (a) (b) SR model ing section [5, 14]. LR model band 7 band 7 band 6 band 6 IV. SHIFT CURRENT IN MONOLAYER SNS A. FIG. 3. Band structures of monolayer SnS obtained from (a) the SR model and (b) the LR model. The SR model captures the essential band gap and symmetry-protected degeneracies along the Y → S → X path, while the LR model improves quantitative accuracy. shift σijk (0; ω, −ω) = − Formalism In this section, we investigate the bulk photovoltaic effect in monolayer SnS, focusing on the shift current contribution. The shift current originates from the realspace displacement of electron wave packets during optical excitation and is governed by the geometric properties of Bloch states. We employ the standard length-gauge formulation of nonlinear optical response developed in Refs. [2, 5, 31]. The detailed derivation of the shift current conductivity is provided in Appendix B. Here, we present the final expression and its physical interpretation. The shift current conductivity is expressed as  i~e3 X X fnm ik kj ij αjk mn (k) Rnm (k) + αmn (k) Rnm (k) , 2 2S E (Emn − ~ω − iη) n,m mn (6) k where Emn = Emk − Enk , fnm = f (Enk ) − f (Emk ), and S is the unit cell area. The transition intensity is defined as j k αjk mn (k) = vmn vnm , (7) ij (k) is the shift vector. and Rnm The shift vector is given by i i ij − ξmm , Rnm (k) = ∂ki φjnm + ξnn (8) where φjnm is the phase of the velocity matrix element j i vnm , and ξmm = ihumk |∂ki |umk i is the Berry connection. Equation (6) provides a transparent physical picture of the shift current. The nonlinear response is determined by the product of the transition intensity αjk mn (k) and ij the shift vector Rnm (k), which represents the real-space displacement of the electron wave packet during optical excitation. In monolayer SnS, the absence of inversion symmetry allows for a finite shift current response. In addition, the crystal symmetry restricts the nonzero components of the shift-current conductivity tensor. For linearly polarized light incident from the z-direction, the symmetry-allowed shift shift shift shift components are σxxx , σxyy , and σyxy = σyyx , whereas the other components vanish. In the following, we analyze these symmetry-allowed components based on the effective tight-binding models. Figure 5 schematically illustrates the shift-current mechanism in monolayer SnS in terms of the transition intensity and shift vector. The transition intensity describes the strength of the optical transition, while the shift vector represents the real-space displacement of the electron wave packet during optical excitation. In the following subsection, we evaluate the symmetry-allowed components using the SR and LR tight-binding models and analyze how the hopping range modifies the shiftcurrent spectra. The comparison between the SR and LR models reveals that the essential features of the shift current are already captured by the SR model, while the LR model provides quantitative corrections. This indicates that the shift current in monolayer SnS is primarily governed by the low-energy electronic structure. These results establish a clear connection between electronic structure, symmetry, and nonlinear optical response, and provide useful insights for designing materials with enhanced bulk photovoltaic effects. B. Shift Current Conductivity in Monolayer SnS We calculate the shift current conductivity of monolayer SnS under linearly polarized light incidence. Monolayer SnS belongs to the P mn21 space group. Due to its symmetry, the shift current conductivity components shift shift shift shift are finite when linearly , σxyy , and σyxy = σyyx σxxx polarized light is incident from the z-direction. All other components vanish. Figure 6 shows the calculated shift current conductivity obtained using both the SR and LR models. It is evident that all three shift current conductivity components exhibit higher peak values in the SR model. 5 (a) (b) small R large R (eV) FIG. 4. Berry curvature and optical conductivity of monolayer SnS. (a) Berry curvature Ωz (k) in the first Brillouin zone calculated using the LR model. The anisotropic distribution reflects inversion symmetry breaking. (b) Longitudinal optical conductivity σxx (ω) for the SR and LR models. The LR model provides quantitative accuracy, while the SR model captures essential features. ij Rnm (k) CB jj αmn VB FIG. 5. Schematic illustration of the shift current mechanism in monolayer SnS. (a) Real-space picture of photoexcited carrier displacement. (b) Interband transition between valence and conduction bands, characterized by the transition intensity and shift vector. optical response, where the shift-vector-related numerator is absent, and highlights the sensitivity of the shiftcurrent response to the detailed momentum-space structure of the electronic states. Therefore, the SR model captures the qualitative low-energy features of the shiftcurrent response, whereas the LR model is necessary for a more quantitatively reliable description. For completeness, Appendix C presents band-resolved ij contour plots of the shift vector Rnm (k) and transition jk j k intensity αmn = vmn vnm for selected interband transitions near the band gap. The band indices are defined by the ascending order of the eigenvalues in the band structure shown in Fig. 3. V. To clarify the origin of the different peak magnitudes between the SR and LR models, we further analyze the momentum-resolved contribution to the shift-current shift conductivity σxyy at the representative low-energy peaks in the left panel of Fig. 6. Figure 7 shows the resulting shift k-resolved contribution to Re[σxyy ] before the Brillouinzone summation, evaluated at ~ω = 2.92 eV for the SR model and ~ω = 2.43 eV for the LR model. The SR model exhibits pronounced local contributions along the resonant regions of the Brillouin zone. In contrast, the corresponding contributions in the LR model are less pronounced and more broadly distributed. The stronger local contributions in the SR model are consistent with the larger peak observed in the shift-current spectrum. This indicates that the SR–LR difference is governed by the momentum-space distribution of the full shiftcurrent integrand, including the transition matrix elements, the shift-vector-related numerator, and the resonance denominator, rather than by the band-energy structure alone. This behavior contrasts with the linear CONCLUSION In this study, we theoretically investigated the electronic structure and optical properties of monolayer SnS. Two effective tight-binding models with different hopping ranges were constructed using maximally localized Wannier functions to reproduce the energy band structure near the band gap. For the first-order optical response, we confirmed that monolayer SnS exhibits finite longitudinal conductivity, validating the SR model. We also demonstrated that Berry curvature originates from inversion symmetry breaking and reverses sign under wavevector inversion, resulting in vanishing transverse conductivity. For the second-order nonlinear optical response, we showed that monolayer SnS generates a shift current under linearly polarized light without external strain, unlike centrosymmetric materials that require strain for BPVE. Furthermore, our simplified SR model successfully captures the essential features of shift current conductivity. These findings provide a transparent framework for understanding shift current responses in monolayer SnS and related non-centrosymmetric 2D materials. 6 2.92 eV large R 2.43 eV large R Re[σyxy] small R Re[σxxx] Re[σxyy] small R (eV) (eV) small R large R (eV) shift shift shift FIG. 6. The real part of shift current conductivities σxyy (left panel), σxxx (center panel) and σyxy (right panel) for monolayer 3 SnS. The unit of shift current conductivity is e /~. (a) SR (b) LR × 104 Y Y 1 X Γ Γ 1st BZ 1st BZ ℏω = 2.92eV X 0 -1 ℏω = 2.43eV shift FIG. 7. Momentum-resolved contribution to Re[σxyy ] at the representative low-energy peaks of Fig. 6. (a) and (b) show the SR and LR models evaluated at ~ω = 2.92 eV and 2.43 eV, respectively. The plotted quantity is the shift-current integrand before Brillouin-zone summation, including the same resonance denominator and broadening factor as in the spectral calculation, with the overall prefactor omitted. ACKNOWLEDGMENTS This work was supported by JSPS KAKENHI (Grants No. JP25K01609, No. JP22H05473, and No. JP21H01019), JST CREST (Grant No. JPMJCR19T1). K. W. acknowledges the financial support for Basic Science Research Projects (Grant No. 2401203) from the Sumitomo Foundation, and the special individual re- search support from Kwansei Gakuin University. T. K. acknowledges the Sasakawa Scientific Research Grant from The Japan Science Society. Appendix A: Details of the Hamiltonian in the SR model Here, we present the Hamiltonian of the short-range (SR) tight-binding model for monolayer SnS. Figure 8 shows the lattice structure. The shaded rhombus denotes the primitive unit cell, which contains four nonequivalent atomic sites labeled A, B, C, and D. The primitive lattice vectors are given by a1 = (4.28, 0) and a2 = (0, 4.08). The position vectors of the atomic sites are defined as rA = (0.473, 3.06), rB = (2.61, 1.02), rC = (2.29, 1.02), and rD = (0.15, 3.06), respectively. In monolayer SnS, the electronic states near the band gap are mainly derived from the p orbitals of Sn and S atoms. Since the unit cell contains two Sn atoms and two S atoms, the effective tight-binding model is constructed from 12 Wannier orbitals, corresponding to three orbitals per atomic site, resulting in a 12 × 12 Hamiltonian. The SR model includes only nearest-neighbor hopping terms. The eigenvalue equation is given by H(k) |ψnk i = Enk |ψnk i , where H(k) is the Bloch Hamiltonian at wavevector k = (kx , ky ), and Enk is the eigenenergy with band index n = 1, 2, . . . , 12. The matrix elements of H(k) are written as ( γαβ eik·ταβ , (α, β) ∈ {(A, D), (D, A), (B, C), (C, B)},  hαβ (k) = γαβ eik·ταβ + e−ik·ταβ , otherwise, where γαβ (in eV) is the hopping parameter between orbitals α and β, as listed in Table I. The relative displace- (A1) (A2) ment vector between sites is defined as ταβ = rβ − rα . This formulation provides a minimal yet physically 7 rD rA rC rB y a2 D A C Sn S B a1 x FIG. 8. Top view of the lattice structure of monolayer SnS. The shaded rhombus indicates the primitive unit cell containing four nonequivalent sites (A–D). transparent model that captures the essential low-energy electronic structure relevant to optical transitions. perturbation Hamiltonian is given by X Ej (t) r̂j , H ′ (t) = e (B1) j where r̂j is j element of the position operator r̂ and Ej (t) = Ej e−iωt + Ej∗ eiωt . The dynamics of the density matrix ρ̂ = ρ̂(0) + ρ̂(1) + ρ̂(2) + · · · are governed by the von Neumann equation Appendix B: Derivation of Shift Current Conductivity In this appendix, we derive the shift current conducshift tivity σijk (0; ω, −ω) within the length-gauge formalism and the velocity-gauge formalism, following Refs. [2, 5, 31]. i~  ∂ ρ̂  = H0 + H ′ (t), ρ̂ . ∂t At equilibrium, ρ̂(0) is given by the Fermi–Dirac distribution. Solving perturbatively, the first-order density matrix element in the band basis is ρ(1) nm (ω) = 1. Second-order optical response We consider the interaction of a crystalline solid with a monochromatic light field in the length gauge. The ′ ′ ρ(2) nm (ω + ω ; ω, ω ) = X e2 ′ Emn − ~(ω + ω ) + iη l where ω and ω ′ are the frequencies of the incident light from j and k directions, respectively. (B2) j e fnm rnm Ej (ω) , Emn − ~ω − iη (B3) where fnm = f (Enk ) − f (Emk ) and j rnm = hunk | r̂ · ej |umk i is the matrix element of the r̂j between states n and m. Similarly, the secondorder density matrix is obtained as k j k j fnl rnl rlm flm rnl rlm − Eml − ~ω + iη Eln − ~ω + iη ! Ej (ω)Ek (ω ′ ), (B4) The DC second-order current is expressed as X (2) (2) Ji = σijk (0; ω, ω ′ ) Ej (ω) Ek (ω ′ ), (B6) jk (2) 2. Shift current conductivity For charge current operator J, the thermal expectation value of J is given as hJi = Tr(Jρ(t)). (B5) where σijk (0; ω, ω ′ ) is the second-order conductivity tensor. In this paper, we focus on the photocurrent generated by BPVE, which corresponds to the case of ω ′ = −ω. We can rewrite the second-order conductivity as (2) injection shift σijk (0; ω, −ω) = σijk (0; ω, −ω) + σijk (0; ω, −ω), (B7) 8 TABLE I. Hopping parameters γαβ of the short-range (SR) tight-binding model for monolayer SnS (in eV). The indices A, B, C, and D denote atomic sites, and the subscripts 1, 2, and 3 correspond to the three p orbitals at each site. A1 -1.927 -0.029 -0.001 -0.118 0.109 -0.032 -1.447 0.641 0.002 2.039 0.092 0.000 A1 A2 A3 B1 B2 B3 C1 C2 C3 D1 D2 D3 A2 -0.029 -0.872 -0.001 -0.078 0.312 -0.057 0.721 2.433 -0.087 -0.320 -0.380 0.000 A3 -0.001 -0.001 0.010 -0.027 0.027 -0.187 -0.024 -0.086 2.892 0.000 0.000 -0.573 B1 -0.118 -0.078 -0.027 -1.927 0.029 -0.001 2.039 -0.092 0.000 -1.457 -0.675 0.012 B2 0.109 0.312 0.027 0.029 -0.872 0.001 0.320 -0.380 0.000 -0.752 2.550 0.074 B3 -0.032 -0.057 -0.187 -0.001 0.001 0.010 0.000 0.000 -0.573 -0.007 0.068 2.838 C1 -1.447 0.721 -0.024 2.039 0.320 0.000 -3.918 0.038 0.000 0.565 -0.191 -0.028 C2 0.641 2.433 -0.086 -0.092 -0.380 0.000 0.038 -4.333 -0.001 0.241 0.407 -0.016 C3 0.002 -0.087 2.892 0.000 0.000 -0.573 0.000 -0.001 -4.170 -0.037 -0.018 0.243 D1 2.039 -0.320 0.000 -1.457 -0.752 -0.007 0.565 0.241 -0.037 -3.918 -0.038 0.000 D2 0.092 -0.380 0.000 -0.675 2.550 0.068 -0.191 0.407 -0.018 -0.038 -4.333 0.001 D3 0.000 0.000 -0.573 0.012 0.074 2.838 -0.028 -0.016 0.243 0.000 0.001 -4.170 injection shift where σijk and σijk are the shift and injection current contributions, respectively. shift Under linearly polarized light irradiation, the optical conductivities of the shift current σijk (0; ω, −ω) and the injection injection current σijk (0; ω, −ω) are given by [2, 32, 33] shift σijk =   e3 X X fnm j j k k , rmn rnm;i + rmn rnm;i 2~S E − ~ω − iη n,m mn (B8) k injection σijk (0; ω, −ω) = τ ie3 X X fnm ∆i {rk , rj }, 2~S E − ~ω − iη nm mn nm mn nm (B9) k j where the generalized derivative rnm;i is expressed as ∂rj j i i j rnm;i = nm − i[ξnn (k) − ξmm (k)]rnm , ∂ki (B10) i where ξmm is the Berry connection for the m-th band, given by i ξmm = i humk | ∂ |umk i . ∂ki (B11) tors:  j  vnm j rnm = iωnm  0 (n 6= m) (B12) (n = m), j j j where vnm = hunk | v̂ · ej |umk i, ∆jmn = vmm − vnn and ωnm = (Enk − Emk )/~. Here, we introduce the shift vector i i ij − ξmm , Rnm (k) = ∂ki φjnm + ξnn (B13) which includes the Berry connection difference. The phase term φjnm corresponds to the phase of the group velocity matrix element as j j j vnm = hunk | v̂ · ej |umk i = |vnm |eφnm . (B14) From Eqs. (B10) and (B13), we obtain 3. Velocity representation and shift vector To evaluate the conductivity numerically, we express the position matrix elements in terms of velocity opera- j ij j rnm;i = −iRnm (k)rnm . (B15) Substituting these expressions into the conductivity formula, Eq. (B8), we obtain the shift current conductivity in terms of the velocity matrix elements and the shift vector: 9 shift σijk =−  i~e3 X X fnm j k ik k j ij . vmn vnm Rnm + vmn vnm Rnm 2 2S E (Emn − ~ω − iη) n,m mn (B16) k Here, we consider velocity representation of the shift ij vector Rnm (k). The generalized derivative of the posij tion operator rnm can be expressed using the sum rule of differentiation, allowing the separation of the BPVE into j rnm;i = shift and injection currents [2, 32, 33]. The generalized j derivative rnm:i is given by i j rnm ∆jmn + rnm ∆imn i X j i i j (ωlm rnl rlm − ωnl rnl rlm ), + ωnm ωnm (B17) l Using Eqs. (B17) and (B10), we obtain ij j −iRnm (k)rnm = i j rnm ∆jmn + rnm ∆imn i X j i i j (ωlm rnl rlm − ωnl rnl rlm ). + ωnm ωnm m (B18) From Eq. (B12), the right-hand side of Eq. (B18) can be rewritten as j j  i i i j j j i i vnl vlm vnl i X vlm ) + vnm (vmm − vnn ) vnm (vmm − vnn ω + − ω lm nl 2 iωnm ωnm iωnl iωlm iωnl iωlm l j i  i j i j j j i i vnl vlm vlm v (v − vnn ) + vnm (vmm − vnn ) i X vnl = nm mm . + + − 2 iωnm ωnm ωnl ωlm l (B19) Thus, we obtain " !# j i i j i j j j i i X vnl vlm vnl vlm vnm (vmm − vnn ) + vnm (vmm − vnn ) +i + − iωnm ωnl ωlm l !# " j i i j i j j j i i vnl vlm vlm (vmm − vnn ) + vnm (vmm − vnn ) X vnl i~ vnm . + − = j Enm Enl Elm vnm 1 ij Rnm (k) = − j vnm (B20) l shift shift shift The shift current conductivity σijk is symmetric with respect to the exchange of j and k. For instance, σyxy = σyyx . For linearly polarized light (j = k), this simplifies to shift =− σijj j ij i~e3 X X fnm |vnm |2 Rnm . 2 S E (Emn − ~ω − iη) n,m mn (B21) k j 2 This expression clearly shows that the shift current is governed by the product of the transition intensity αjj mn = |vnm | ij and the shift vector Rnm (k), which represents the real-space displacement of the electron wave packet during optical excitation. This geometric nature underlies the bulk photovoltaic effect [5]. 4. Electronic Polarization and Berry Connection The electronic polarization can be expressed in terms of the Berry connection as Z dk X i Pi = e fm ξmm , (B22) 3 BZ (2π) m where the Berry connection represents the Wannier center of wavefunctions in real space. This means that the real-space electric polarization is determined by the momentum-space Berry connection. From this perspective, shift current originates from the difference in the charge center between the valence and conduction bands upon optical excitation. However, shift current does not generally appear in all systems: it vanishes in centrosymmetric systems. Even in noncentrosymmetric systems, shift the allowed tensor components of σijk (0; ω, −ω) are restricted by the crystal symmetry. 10 (C1) shift This function determines whether σijk is finite. When it shift is an even function with respect to kx = 0, σijk is finite, shift whereas when it is an odd function, σijk vanishes. ij For j = k, the function simplifies to Rnm (k)αjj mn (k). xx xx For example, in Fig. 9, R67 (k) and α76 (k) are both shift even functions, ensuring that σxxx is finite. Similarly, xy yy R67 (k) and α76 (k) are both even functions, confirming yy shift that σxyy is finite, whereas for R67 (k)αyy 76 (k), the product of an odd and even function is an odd function, leadshift shift ing to σyyy = 0. For σyyx , we examine the function yx yx xy yy α76 (k)R67 (k) + α76 (k)R67 (k). Since both αyx 76 (k) and yx R67 (k) are odd functions, their product is even. Simyy ilarly, αxy 76 (k) and R67 (k) are even functions, making their product even. Since the sum of these two terms shift is nonzero, σyyx is finite. The same analysis can be applied to other conductivity components, yielding similar results for the LR model as shown in Fig. 10. [1] V. M. Fridkin, Bulk photovoltaic effect in noncentrosymmetric crystals, Crystallography Reports 46, 654 (2001). [2] J. E. Sipe and A. I. Shkrebtii, Second-order optical response in semiconductors, Phys. Rev. B 61, 5337 (2000). [3] S. M. Young and A. M. Rappe, First principles calculation of the shift current photovoltaic effect in ferroelectrics, Phys. Rev. Lett. 109, 116601 (2012). [4] T. Morimoto and N. Nagaosa, Topological nature of nonlinear optical effects in solids, Science Advances 2, e1501524 (2016). [5] T. Morimoto, S. Zhong, J. Orenstein, and J. E. Moore, Semiclassical theory of nonlinear magneto-optical responses with applications to topological dirac/weyl semimetals, Phys. Rev. B 94, 245121 (2016). [6] B. M. Fregoso, T. Morimoto, and J. E. Moore, Quantitative relationship between polarization differences and the zone-averaged shift photocurrent, Phys. Rev. B 96, 075421 (2017). [7] W. Shockley and H. J. Queisser, Detailed balance limit of efficiency of p-n junction solar cells, J. Appl. Phys. 32, 510 (1961). [8] J. E. Spanier, V. M. Fridkin, A. M. Rappe, A. R. Akbashev, A. Polemi, Y. Qi, Z. Gu, S. M. Young, C. J. Hawley, D. Imbrenda, et al., Power conversion efficiency exceeding the shockley–queisser limit in a ferroelectric insulator, Nature Photonics 10, 611 (2016). [9] A. M. Cook, B. M. Fregoso, F. de Juan, S. Coh, and J. E. Moore, Design principles for shift current photovoltaics, Nat. Commun. 8, 14176 (2017). [10] M. Wu and X. C. Zeng, Intrinsic ferroelasticity and/or multiferroicity in two-dimensional phosphorene and phosphorene analogues, Nano Lett. 16, 3236 (2016). [11] W. Ding, J. Zhu, Z. Wang, Y. Gao, D. Xiao, Y. Gu, Z. Zhang, and W. Zhu, Prediction of intrinsic twodimensional ferroelectrics in in2se3 and other iii2-vi3 van der waals materials, Nat. Commun. 8, 14956 (2017). [12] R. Fei, W. Kang, and L. Yang, Ferroelectricity and phase transitions in monolayer group-iv monochalcogenides, Phys. Rev. Lett. 117, 097601 (2016). [13] J. Qiao, X. Kong, Z.-X. Hu, F. Yang, and W. Ji, High-mobility transport anisotropy and linear dichroism in few-layer black phosphorus, Nat. Commun. 5, 4475 (2014). [14] T. Rangel, B. M. Fregoso, B. S. Mendoza, T. Morimoto, J. E. Moore, and J. B. Neaton, Large bulk photovoltaic effect and spontaneous polarization of single-layer monochalcogenides, Phys. Rev. Lett. 119, 067402 (2017). [15] N. Higashitarumizu, H. Kawamoto, C.-J. Lee, B.-H. Lin, F.-H. Chu, I. Yonemori, T. Nishimura, K. Wakabayashi, W.-H. Chang, and K. Nagashio, Purely in-plane ferroelectricity in monolayer sns at room temperature, Nat. Commun. 11, 2428 (2020). [16] H. Kawamoto, N. Higashitarumizu, N. Nagamura, M. Nakamura, K. Shimamura, N. Ohashi, and K. Nagashio, Micrometer-scale monolayer sns growth by physical vapor deposition, Nanoscale 12, 23274 (2020). [17] Y. Bao, P. Song, Y. Liu, Z. Chen, M. Zhu, I. Abdelwahab, J. Su, W. Fu, X. Chi, W. Yu, et al., Gate-tunable in-plane ferroelectricity in few-layer sns, Nano Lett. 19, 5109 (2019). [18] L. Z. Tan, S. M. Young, F. Zheng, and A. M. Rappe, Shift current bulk photovoltaic effect in polar materials—hybrid and oxide perovskites and beyond, npj Computational Materials 2, 16026 (2016). [19] R. Fei, W. Kang, and L. Yang, Ferroelectricity and phase transitions in monolayer group-iv monochalcogenides, Phys. Rev. Lett. 117, 097601 (2016). [20] G. Kresse and J. Furthmüller, Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set, Phys. Rev. B 54, 11169 (1996). [21] G. Kresse and D. Joubert, From ultrasoft pseudopotentials to the projector augmented-wave method, Phys. Rev. B 59, 1758 (1999). [22] P. E. Blöchl, Projector augmented-wave method, Phys. Rev. B 50, 17953 (1994). [23] J. P. Perdew, K. Burke, and M. Ernzerhof, Generalized gradient approximation made simple, Phys. Rev. Lett. 77, 3865 (1996). [24] H. J. Monkhorst and J. D. Pack, Special points for brillouin-zone integrations, Phys. Rev. B 13, 5188 (1976). Appendix C: Contour plots of shift vector and transition intensity By constructing the effective model, Here, we show ij contour plots of the shift vector Rnm (k) and the tranjk j k sition intensity αmn = vmn vnm . This facilitates an intuitive understanding of the shift current conductivity shift σijk , analogous to the relationship between the Berry curvature and the transverse conductivity in linear optical responses. ij Figures 9 and 10 show contour plots of Rnm (k) and j k αjk (k) = v v for n = 6, m = 7 in the SR and LR mn mn nm models, respectively. In Eq. (6), the non-integral function is expressed as the product ik kj ij αjk mn (k)Rnm (k) + αmn (k)Rnm (k). 11 SR model xx R67 yy R67 xy R67 yx R67 xx α76 yy α76 xy α76 yx α76 ij j k FIG. 9. Contour plot of shift vector Rnm (k) and transition intensity αjk mn = vmn vnm in the SR model. xx R67 yy R67 xx α76 yy α76 LR model xy R67 yx R67 xy α76 yx α76 ij j k FIG. 10. Contour plot of shift vector Rnm (k) and transition intensity αjk mn = vmn vnm in the LR model. [25] N. Marzari and D. Vanderbilt, Maximally localized generalized wannier functions for composite energy bands, Phys. Rev. B 56, 12847 (1997). [26] I. Souza, N. Marzari, and D. Vanderbilt, Maximally localized wannier functions for entangled energy bands, Phys. Rev. B 65, 035109 (2001). [27] N. Marzari, A. A. Mostofi, J. R. Yates, I. Souza, and D. Vanderbilt, Maximally localized wannier functions: Theory and applications, Rev. Mod. Phys. 84, 1419 (2012). [28] A. A. Mostofi, J. R. Yates, Y.-S. Lee, I. Souza, D. Vanderbilt, and N. Marzari, wannier90: A tool for obtaining maximally-localised wannier functions, Comput. Phys. Commun. 178, 685 (2008). [29] R. Kubo, Statistical-mechanical theory of irreversible processes. i. general theory and simple applications to magnetic and conduction problems, J. Phys. Soc. Jpn. 12, 570 (1957). 12 [30] D. Xiao, M.-C. Chang, and Q. Niu, Berry phase effects on electronic properties, Rev. Mod. Phys. 82, 1959 (2010). [31] C. Aversa and J. E. Sipe, Nonlinear optical susceptibilities of semiconductors: Results with a length-gauge analysis, Phys. Rev. B 52, 14636 (1995). [32] R. Habara and K. Wakabayashi, Symmetry manipulation of nonlinear optical effect for metallic transition-metal dichalcogenides, Phys. Rev. B 107, 115422 (2023). [33] H. Xu, H. Wang, J. Zhou, and J. Li, Pure spin photocurrent in non-centrosymmetric crystals: bulk spin photovoltaic effect, Nat. Commun. 12, 4330 (2021).