Regularized lattice theory for spatially dispersive nonlinear optical conductivities

Steven Gassner,Eugene J. Mele
DOI: https://doi.org/10.1103/PhysRevB.108.085403
2023-04-26
Abstract:Nonlinear optical responses are becoming increasingly relevant for characterizing the symmetries and quantum geometry of electronic phases in materials. Here, we develop an expanded diagrammatic scheme for calculating spatially dispersive corrections to nonlinear optical conductivities, which we expect to enhance or even dominate even-order responses in materials of recent interest. Building upon previous work that enforces gauge invariance of spatially uniform nonlinear optical responses, we review the cancellation of diagrams required to ensure the equivalence between velocity gauge and length gauge formulations, and provide a simple vertex rule for extending optical responses to first order in the light wave vector q. We then demonstrate the method with calculations on a prototypical centrosymmetric model where spatial dispersion admits anomalous secondharmonic generation, a response that is symmetry-forbidden under the dipole approximation.
Mesoscale and Nanoscale Physics,Optics
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: how to calculate the spatial dispersion correction of the nonlinear optical response in materials, especially how these corrections enhance or dominate the even - order response in materials of recent interest. Specifically, the author focuses on how to calculate the nonlinear optical conductivity through an extended diagrammatic scheme in the case of considering the spatial variation of the electromagnetic field, especially in the case of the first - order light - wave vector \(q\). ### Specific background and motivation of the problem 1. **Importance of nonlinear optical response** - The nonlinear optical response is crucial for characterizing the symmetry and quantum - geometric properties of materials. - These responses are also of great significance in optoelectronic, laser and other devices. 2. **Limitations of existing methods** - Traditional nonlinear optical calculations are usually based on the dipole approximation, which may be inaccurate in some cases, especially in the band structure where large - scale coherence is introduced by artificial lattices or Wannier representations. - The existing theoretical frameworks have difficulties in dealing with the equivalence between the velocity gauge and the length gauge, especially when the effective Hamiltonian is projected onto a partial Hilbert space. 3. **Research objectives** - Develop an extended diagrammatic scheme for calculating the spatial dispersion correction of the nonlinear optical conductivity. - Ensure the equivalence between the velocity gauge and the length gauge, and provide a simple vertex rule to extend the optical response to the first - order light - wave vector \(q\). - Verify the effectiveness of this method through specific models, especially by showing the anomalous second - harmonic - generation (SHG) response that occurs in centrosymmetric systems. ### Main contributions of the paper - **Extended diagrammatic scheme**: A new diagrammatic method is proposed, which can systematically calculate the spatial dispersion correction of the nonlinear optical conductivity. - **Vertex rule**: A new vertex rule is introduced to handle the first - order correction of the light - wave vector \(q\) in the velocity gauge. - **Model verification**: Through a typical centrosymmetric model, the effectiveness of this method in calculating the anomalous second - harmonic - generation response is demonstrated. - **Physical significance**: Explain why a non - zero second - harmonic - generation response can be observed even in centrosymmetric systems when considering the spatial variation of the electromagnetic field. ### Formula summary - **Hamiltonian in the velocity gauge** \[ \hat{H}_k(t)=\hat{H}_k^{(0)}+e\mathbf{A}(t)\cdot\nabla_k\hat{H}_k^{(0)}+\frac{e^2}{2}\mathbf{A}(t)\cdot\nabla_k\left(\mathbf{A}(t)\cdot\nabla_k\hat{H}_k^{(0)}\right)+\cdots \] - **Density matrix equation** \[ i\hbar\partial_t\hat{\rho}(t)=[\hat{H}(t),\hat{\rho}(t)] \] - **Nonlinear conductivity tensor** \[ \sigma_{\mu\alpha_1\ldots\alpha_n}(\omega;\omega_1,\ldots,\omega_n)=-\frac{e^{n + 1}}{\hbar^n}\prod_{j = 1}^n\left(\frac{1}{i\omega_j}\right)\int [dk]\kappa_{\mu\alpha_1\ldots\alpha_n}^k(\omega;\{\omega_j\}) \] - **Spatial dispersion correction** \[ \sigma_{\mu\alpha_1\ldots\alpha_n}(q)=\sigma_{\mu\alpha_1\ldot