Existence of a phase with finite localization length in the double scaling limit of N-orbital models
Section snippets
The site-diagonal-disorder model
In the first sections of this article we study Wegner’s site-diagonal-disorder model with orbitals on each site, which at reduces to the Anderson model [1], [2]. The on-site Hamiltonian for site is an random matrix which describes interactions between the orbitals at site . It is the natural generalization of the Anderson model’s disorder potential. The mean squared value of ’s matrix elements is proportional to , so that when a single site is taken in isolation, its
Numerical calculation of the localization length when the hopping strength is kept fixed
In Fig. 1 we present the localization length , which we calculate using the transfer matrix method [15]. The transfer matrix method for calculating localization lengths operates by building a long chain with length and measuring the average decay within the chain. We keep the energy fixed at , fairly close to the band edge which is near in the 1-D chain, in order to reduce the scattering length and the computational cost.
We calculate with the hopping strength fixed at three
Numerical calculation of the localization length in the double scaling limit where is kept fixed
Now we turn to the more interesting case where the product is kept fixed while taking the limit. Fig. 2 shows the localization length with fixed at for both the unitary and orthogonal ensembles. When is fixed to these values the scattering length always remains small and does not significantly affect the computational cost of calculating the localization length. Therefore we fix the energy at the band center and calculate at large values of up to .
Fig. 2
Discussion of the numerical results
These results are a numerical proof that the one-dimensional site-diagonal-disorder model is localized in the double scaling limit where is kept fixed and is taken to infinity. This localized property is expected of one-dimensional wires with short-range uncorrelated disorder and short-range hopping. Wires in this class have been shown both numerically and analytically to be localized, except in the special case of certain ensembles with perfectly conducting channels [4], [5].
These
The gauge invariant model
In order to go beyond one dimension, and to obtain non-numerical insight, we will now analyze both the fixed limit and the fixed limit using analytical analysis of a functional integral. In order to ease our mathematical analysis, we will switch from Wegner’s site-diagonal-disorder model to Wegner’s gauge invariant model, both of which were introduced in the same paper [1]. The only difference between the two models is that in the gauge invariant model the hopping between sites is mediated
Functional integral formulation of the gauge invariant model
The conversion of disordered models to functional integrals is a well developed topic that originated with Schafer and Wegner in 1980 [19], [20]. The main point of these functional integrals is that their spin variables obey a continuous symmetry and therefore are capable of a transition from a spontaneously-broken-symmetry phase to a symmetric phase. In the symmetry-broken phase the spins are correlated over long distances, which corresponds to electronic conduction over long distances. In the
Two energy and length scales
In this section we will use the hybrid functional integral in Eq. (6) to show that the gauge invariant model has two natural energy scales. The first is the scattering energy scale . The second is the much smaller energy scale which controls conduction and localization. Corresponding to these two energy scales are two length scales: the scattering length which is equal to 1 in the gauge invariant model because scattering occurs on site, and the localization length which can be much
Summary
In summary, we have presented evidences that two disordered models, each with orbitals per site, possess a localized phase with finite localization length in the double scaling limit where and is kept fixed. In the site-diagonal-disorder model we directly computed the localization length and showed that it remains finite in one dimension when is kept fixed. In the gauge invariant model, in any dimension or geometry, we showed that the fixed limit preserves dynamic spin variables
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
We especially thank Università degli Studi di Roma “La Sapienza”, Italy, the Asia Pacific Center for Theoretical Physics, South Korea, Sophia University, and the Zhejiang Institute of Modern Physics, where much of this work was done. We thank T. Ohtsuki for help during the early stages of this work. We also acknowledge support from the Institute of Physics of the Chinese Academy of Sciences, Nanyang Technological University, Singapore, Royal Holloway University of London, United Kingdom, and
References (48)
- et al.
Ann. Phys.
(1994) Nuclear Phys. B
(2002)- et al.
Nuclear Phys. B
(2002) - et al.
Nuclear Phys. B
(1993) Phys. Rev. B
(1979)Phys. Rev.
(1958)- et al.
Comm. Math. Phys.
(1993) Progr. Theoret. Phys. Suppl.
(1973)- et al.
Rep. Progr. Phys.
(1993) Phys. Rev. Lett.
(1992)
J. Phys. Soc. Japan
J. Phys. Soc. Japan
Phys. Rev. Lett.
Phys. Rev. Lett.
Eigenvalue Statistics and Localization for Random Band Matrices with Fixed Width and Wegner Orbital Model
Int. Math. Res. Not.
Phys. Rev. Lett.
Z. Phys. B
Phys. Rev. Lett.
J. Phys. A: Math. Gen.
Sov. Phys.—JETP
Supersymmetry in Disorder and Chaos
Cited by (1)
Special Issue “Localisation 2020”: Editorial Summary
2021, Annals of Physics