journal article·2026-03-05

Valley physics in the two-band k · p model for SiGe heterostructures and spin qubits

Tancredi Salamone, Biel Martinez Diaz, Jing Li, Lukas Cvitkovich, Yann-Michel Niquet

Physical Review B· Volume 113· Issue 11
Publisher page

Abstract

We discuss the choice and implementation of inter-valley potentials in the so-called two bands $\mathbf{k}\cdot\mathbf{p}$ model for the opposite $X$, $Y$ or $Z$ valleys of silicon. We focus on the description of valley splittings in Si/SiGe heterostructures for spin qubits, with a particular attention to alloy disorder. We demonstrate that the two bands $\mathbf{k}\cdot\mathbf{p}$ model reproduces the valley splittings of atomistic tight-binding calculations in relevant heterostructures (SiGe spikes, wiggle wells...), yet at a much lower cost. We show that the model also captures the effects of valley-orbit mixing and yields the correct inter-valley dipole matrix elements that characterize manipulation, dephasing and relaxation in spin/valley qubits. We simulate a realistic Si/SiGe spin qubit device as an illustration, and discuss electron-phonon interactions in the two bands $\mathbf{k}\cdot\mathbf{p}$ model. Beyond spin qubits, this model enables efficient simulations of SiGe heterostructure devices where spin and valley physics are relevant.
Source: openalex_inverted_index · Quality: baseline

Research context

OpenAlex topics: Quantum and electron transport phenomena · Semiconductor Quantum Structures and Devices · Topological Materials and Phenomena

SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.

Affiliations

Université Grenoble Alpes

CEA

Leti

University of Regensburg

IRIG-MEM-L_Sim

Funding

Agence Nationale de la Recherche · HORIZON EUROPE Framework Programme · Grand Équipement National De Calcul Intensif

References (100+)

Show source-provided references
1. 10.1103/RevModPhys.95.02500310.1103/RevModPhys.95.025003 ↗
2. 10.1038/s41534-017-0038-y10.1038/s41534-017-0038-y ↗
3. 10.1038/s41586-022-04986-610.1038/s41586-022-04986-6 ↗
4. 10.1103/PhysRevX.14.01104810.1103/PhysRevX.14.011048 ↗
5. 10.1038/ncomms1357510.1038/ncomms13575 ↗
6. 10.1038/s41928-022-00722-010.1038/s41928-022-00722-0 ↗
7. 10.1038/s41534-023-00776-810.1038/s41534-023-00776-8 ↗
8. 10.1038/s41567-024-02481-510.1038/s41567-024-02481-5 ↗
9. 10.1038/s41586-025-09531-910.1038/s41586-025-09531-9 ↗
10. 10.1038/s41586-022-05117-x10.1038/s41586-022-05117-x ↗
11. 10.1038/s41586-021-04182-y10.1038/s41586-021-04182-y ↗
12. 10.1126/sciadv.abn513010.1126/sciadv.abn5130 ↗
13. 10.1038/s41586-021-04273-w10.1038/s41586-021-04273-w ↗
14. 10.1038/s41586-023-05777-310.1038/s41586-023-05777-3 ↗
15. 10.1063/5.016084710.1063/5.0160847 ↗
16. 10.1021/acs.nanolett.4c0520510.1021/acs.nanolett.4c05205 ↗
17. 10.1038/s41586-021-03332-610.1038/s41586-021-03332-6 ↗
18. 10.1038/s41467-025-65577-310.1038/s41467-025-65577-3 ↗
19. 10.1126/science.ado591510.1126/science.ado5915 ↗
20. 10.1038/s41565-024-01817-910.1038/s41565-024-01817-9 ↗
21. 10.1002/adfm.20180761310.1002/adfm.201807613 ↗
22. 10.1063/5.000201310.1063/5.0002013 ↗
23. 10.1103/PhysRevApplied.17.02402210.1103/PhysRevApplied.17.024022 ↗
24. 10.1103/PhysRevApplied.22.02403010.1103/PhysRevApplied.22.024030 ↗
25. 10.1103/mtky-93p110.1103/mtky-93p1 ↗
26. 10.1103/RevModPhys.85.96110.1103/RevModPhys.85.961 ↗
27. 10.1063/1.356971710.1063/1.3569717 ↗
28. 10.1038/ncomms306910.1038/ncomms3069 ↗
29. 10.1063/1.497251410.1063/1.4972514 ↗
30. 10.1103/PhysRevB.95.16542910.1103/PhysRevB.95.165429 ↗
31. 10.1103/PhysRevLett.119.17680310.1103/PhysRevLett.119.176803 ↗
32. 10.1063/1.503344710.1063/1.5033447 ↗
33. 10.1038/s41534-018-0075-110.1038/s41534-018-0075-1 ↗
34. 10.1103/PhysRevApplied.13.03406810.1103/PhysRevApplied.13.034068 ↗
35. 10.1103/PhysRevApplied.15.04403310.1103/PhysRevApplied.15.044033 ↗
36. 10.1103/PhysRevApplied.17.03404710.1103/PhysRevApplied.17.034047 ↗
37. 10.1038/s41534-024-00826-910.1038/s41534-024-00826-9 ↗
38. 10.1038/s41534-024-00852-710.1038/s41534-024-00852-7 ↗
39. 10.1038/s41534-024-00827-810.1038/s41534-024-00827-8 ↗
40. 10.1103/PRXQuantum.4.02030510.1103/PRXQuantum.4.020305 ↗
41. 10.1103/PRXQuantum.5.04032210.1103/PRXQuantum.5.040322 ↗
42. 10.1103/PhysRevLett.94.18640310.1103/PhysRevLett.94.186403 ↗
43. 10.1063/1.238797510.1063/1.2387975 ↗
44. 10.1038/nphys47510.1038/nphys475 ↗
45. 10.1103/PhysRevB.75.11531810.1103/PhysRevB.75.115318 ↗
46. 10.1103/PhysRevB.80.08130510.1103/PhysRevB.80.081305 ↗
47. 10.1103/PhysRevB.84.15532010.1103/PhysRevB.84.155320 ↗
48. 10.1103/PhysRevB.104.08540610.1103/PhysRevB.104.085406 ↗
49. 10.1021/acs.nanolett.4c0532610.1021/acs.nanolett.4c05326 ↗
50. 10.1103/PhysRevB.106.08530410.1103/PhysRevB.106.085304 ↗
51. 10.1038/s41467-022-35510-z10.1038/s41467-022-35510-z ↗
52. 10.1038/s41467-022-35458-010.1038/s41467-022-35458-0 ↗
53. 10.1103/PhysRevB.108.12540510.1103/PhysRevB.108.125405 ↗
54. 10.1088/2633-4356/acd74310.1088/2633-4356/acd743 ↗
55. 10.1103/frx3-41bz10.1103/frx3-41bz ↗
56. 10.1063/1.163771810.1063/1.1637718 ↗
57. 10.1103/PhysRevB.73.23533410.1103/PhysRevB.73.235334 ↗
58. 10.1109/TED.2007.90287910.1109/TED.2007.902879 ↗
59. 10.1109/TED.2007.90487710.1109/TED.2007.904877 ↗
60. 10.1103/PhysRevB.77.19331110.1103/PhysRevB.77.193311 ↗
61. 10.1103/PhysRevB.98.16543810.1103/PhysRevB.98.165438 ↗
62. 10.1103/PhysRevB.82.20531510.1103/PhysRevB.82.205315 ↗
63. 10.1103/PhysRevMaterials.8.03620210.1103/PhysRevMaterials.8.036202 ↗
64. 10.1103/4sdz-f9cr10.1103/4sdz-f9cr ↗
65. 10.1103/PhysRevB.81.11532410.1103/PhysRevB.81.115324 ↗
66. 10.1103/PhysRevB.88.03531010.1103/PhysRevB.88.035310 ↗
67. 10.1103/PhysRev.138.A22510.1103/PhysRev.138.A225 ↗
68. V. Sverdlov · 2007
69. 10.1016/j.sse.2008.06.01910.1016/j.sse.2008.06.019 ↗
70. 10.1016/j.sse.2013.02.05510.1016/j.sse.2013.02.055 ↗
71. 10.1016/j.sse.2015.02.00710.1016/j.sse.2015.02.007 ↗
72. 10.1016/j.physrep.2015.05.00210.1016/j.physrep.2015.05.002 ↗
73. C. Mattiussi · 200010.1016/S1076-5670(00)80012-9 ↗
74. The Gz−Gz′=4π/a term is proportional to 〈ũ+|ũ−〉Ω0, where |ũ±〉 are defined by Eq. ( (10b) and can be mapped (up to phase factors) onto the Bloch functions |û±〉 at Z (see Sec. 2b). Since these Bloch functions are orthogonal by design, the Gz−Gz′=4π/a term is zero.
75. Note that the phase of A depends on the choice of phase for the Bloch functions. This (global) phase is however irrelevant as long as all intervalley matrix elements are computed consistently.
76. Note that Eq. ( (8) is not strictly invariant by a rigid shift of the potential V(r) (since the squared envelope function may also have small 2k0 components), although shifts of the order of the 100 meV only change the valley splitting by 2%–3% [as tested with the potential of Eq. ( (24)]. This results from the fact that the valley wave functions defined by Eq. ( (1) are not strictly orthogonal.
77. We point out that this (Shannon-Nyquist) constraint on the Fourier transform of f(r) also applies to tight-binding models, which effectively sample potentials on a grid with atomic resolution.
78. 10.1021/nl301099510.1021/nl3010995 ↗
79. 10.1038/s41534-024-00853-610.1038/s41534-024-00853-6 ↗
80. 10.1103/PhysRevLett.133.03700110.1103/PhysRevLett.133.037001 ↗
81. 10.1103/PhysRevLett.107.10720310.1103/PhysRevLett.107.107203 ↗
82. 10.1103/PhysRevB.69.11533310.1103/PhysRevB.69.115333 ↗
83. 10.1103/PhysRevB.98.24542410.1103/PhysRevB.98.245424 ↗
84. 10.1103/PhysRevX.9.02102810.1103/PhysRevX.9.021028 ↗
85. 10.1103/PhysRevB.79.24520110.1103/PhysRevB.79.245201 ↗
86. 10.1103/PhysRevB.59.175810.1103/PhysRevB.59.1758 ↗
87. 10.1103/PhysRevB.50.1795310.1103/PhysRevB.50.17953 ↗
88. 10.1063/1.240466310.1063/1.2404663 ↗
89. 10.1088/0268-1242/12/12/00110.1088/0268-1242/12/12/001 ↗
90. 10.1088/0268-1242/13/11/00210.1088/0268-1242/13/11/002 ↗
91. The Δso's were actually computed with the generalized gradients approximation (GGA). We have compared vasp with the all-electron ab initio code fleur [111] in the local density approximation (LDA) and GGA. All values are in the 1.70–1.85meV nm range.
92. 10.1103/PhysRevLett.131.09700210.1103/PhysRevLett.131.097002 ↗
93. As discussed in Appendix pp1, minimal TB models with only one orbital per atom [59] indeed yield |A|=1, thus, Vinter(r)=VCBO(r). This is, however, supported neither by multiorbital TB models nor by ab initio calculations [50].
94. Nk actually follows a binomial distribution B[n=8Ωk/a3,p=Y(zk)]. We have made the approximation Nk≈8Y(zk)Ωk/a3 valid when the area of the simulation box is much larger than the lateral extent of the wave functions.
95. We use symmetric operator ordering γikαkβ→−12(∂∂αγi∂∂β+∂∂βγi∂∂α) where the effective masses depend on position.
96. The electric field is opposite in the well and barrier (sawtooth potential) in order to meet periodic boundary conditions over the supercell.
97. 10.1103/PhysRev.145.63710.1103/PhysRev.145.637 ↗
98. Namely, ɛzz=−2c12/c11ɛ∥, where c11 and c12 are the bulk elastic constants, and ɛxx=ɛyy=ɛ∥=(a−ab)/a are the in-plane strains, with a the lattice parameter of the Si1−YGeY material at a given z and ab the lattice parameter of the buffer. The elastic constants and lattice parameter of Si1−YGeY are linearly interpolated between those of pure Si and Ge.
99. 10.1021/acs.nanolett.5c0284810.1021/acs.nanolett.5c02848 ↗
100. 10.1063/1.492129110.1063/1.4921291 ↗