Improvement of the Simmons model for tunnel junctions

I. M. W. Räisänen, I. J. Maasilta · 2025

Read Improvement of the Simmons model for tunnel junctions by I. M. W. Räisänen, I. J. Maasilta (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-16337.txt.

Introduction

Ilmo Räisänen and Ilari Maasilta improve the Simmons model, the standard formula for electric current tunneling through a thin insulating barrier in a metal junction. Fitting the model to measurements is a common way to estimate barrier thickness and height, so formula errors flow straight into those estimates.

The authors derive new analytic formulas for current density and conductance that hold at finite voltage and temperature. Their conductance curves sit much closer to the more exact Wentzel-Kramers-Brillouin approximation than the Simmons formulas or the popular parabolic shortcut. Fitting real junction data with the improved model gives noticeably different barrier values, showing the correction matters in practice, not just in theory.

Catalog record

CREATORI. M. W. Räisänen, I. J. Maasilta
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE79 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-16337

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Improvement of the Simmons model for tunnel junctions Preprint: AIP/123-QED I. M. W. Räisänen, I. J. Maasilta Abstract The Simmons model is a well-known and widely used model for the elastic tunneling current of a metallic tunnel junction, and fitting it to electrical measurements can be used to estimate thicknesses and heights of the tunnel barriers. We present here an improvement of the Simmons model, deriving new more accurate analytical formulas for the tunneling current density and conductance at finite voltage and temperature. We demonstrate that our conductance-voltage formulas are much closer to the Wentzel-Kramers-Brillouin approximation than the Simmons model and its commonly used simplified parabolic approximation. In addition, we demonstrate the practical use of our model, by fitting it to experimental tunnel junction conductance-voltage data and showing a sizeable difference from the Simmons model. I Introduction Tunnel junctions 31 , i.e., two electrical conductors separated by a thin electrically insulating barrier allowing quantum mechanical electron tunneling, are versatile and highly useful devices in many application areas of physics. In particular, superconducting tunnel junctions have in recent years been widely applied as components of various advanced devices, such as building blocks of superconducting quantum bits 14 , superconducting quantum interference devices (SQUIDs) 5 , sensitive radiation detectors 7 , on-chip electronic coolers and thermometers 10 ; 18 and metrological applications for the definition of Ampere 22 . This pull from applications has generated renewed interest from the materials science perspective to explore more advanced barrier materials and fabrication methods 20 ; 25 , in addition to the most commonly used thermally grown amorphous aluminum oxide barriers 31 . With such experimental advances, it is also important to critically review established theories, as they describe how various material parameters affect the tunneling current. We limit the discussion here to the case of metal-insulator-metal tunnel junctions, for the sake of simplicity. From the barrier perspective, junctions with one or two superconducting electrodes can always be measured as normal metal junctions above the superconducting critical temperature, to avoid the strong extra features in the tunneling current originating from superconductivity 31 . In most cases, it is enough to work with the Wentzel-Kramers-Brillouin (WKB) approximation 31 ; 3 , which, however, requires numerical integration even for the simplest barrier shapes, and is therefore not very practical when trying to determine the barrier properties from experimental data. For that reason, approximate analytical formulas describing the elastic tunneling current as a function of bias voltage and temperature were developed in the past, most notably by Simmons 28 ; 27 ; 26 ; 29 . His formulas were derived for a general barrier shape, the main parameters being the barrier thickness and the average barrier height. The results by Simmons demonstrate that the barrier parameters not only influence the value of the zero-bias tunneling conductance, but also the rate at which the junction conductance increases with bias voltage and temperature. In other words, the tunneling current is not a purely linear (Ohmic) function of the voltage. This allows fitting of the conductance-voltage (G-V) curves to determine both barrier height and thickness independently. For that reason, Simmons’ model and its simplified low-voltage parabolic conductance-voltage (G-V) expansion have been widely used to estimate thicknesses and heights of tunnel barriers 15 ; 9 ; 16 ; 21 ; 11 ; 6 ; 24 ; 4 ; 30 ; 1 ; 2 . Another widely used approximative analytical model for zero-temperature conductance was derived by Brinkman, Dynes and Rowell (BDR) 3 for the case of an asymmetric trapezoidal barrier, giving identical results to the Simmons model for the curvature and the value of the conductance minimum in the lowest order. In this paper, we revisit the theory for metal-insulator-metal tunnel junctions within the WKB approximation and derive new, more accurate analytical expressions for tunneling current and conductance as a function of both bias voltage and temperature. We demonstrate that our expressions approximate the numerical WKB results much more closely than the old expressions in the literature, using realistic barrier parameters. We also derive expressions for the low-voltage parabolic approximation for the conductance-voltage G(V) curves within our model, including the temperature dependence, which can easily be used to fit experimental data. As an example, we fit a set of experimental G-V data, demonstrating a sizeable difference to the widely used parabolic approximation of the Simmons model, even for the most common barrier material of thermally grown aluminum oxide. II Tunneling in the WKB approximation Let us consider a tunnel junction with a general tunnel barrier, illustrated in Fig. 1 , with a barrier \phi(x) depending on the position x along the direction of current flow. Here \phi(x) is measured from the chemical potential \mu of the negatively biased electrode (left electrode in Fig. 1 ). According to the WKB approximation, the probability that an electron whose kinetic energy in the x direction is E_{x} can tunnel through the barrier is given by 31 D(E_{x})=\exp\left[-\frac{4\pi\sqrt{2m{{}^{*}}}}{h}\int_{x_{1}}^{x_{2}}(\mu+\phi(x)-E_{x})^{\frac{1}{2}}dx\right], (1) where h is the Planck constant and m^{*} is the effective mass of an electron in the barrier. The integration limits x_{1} and x_{2} are the classical turning points at energy E_{x} , which in general are thus functions of E_{x} for a potential that does not have infinitely sharp edges. This energy dependence of the tunneling path length x_{2}-x_{1} is typically not taken into account, and Simmons and others 27 ; 31 approximate the turning points to be constants s_{1} and s_{2} , defined as the turning points at energy \mu (Fig. 1 ). This approximation is exact for the case of rectangular and trapezoidal barriers, for which the edges are sharply defined and therefore the classical turning points are constants. Figure 1: A general tunnel barrier formed by an insulating film between two metal electrodes. Electrode 1 with chemical potential \mu_{1}=\mu is negatively biased with voltage V with respect to electrode 2 that has a chemical potential \mu_{2}=\mu-eV shifted by the bias voltage V . The net flow of electrons takes place from electrode 1 to electrode 2. With D(E_{x}) known, the elastic tunneling current density J as a function of voltage V and temperature T can then be calculated from 27 ; 31 J(V,T)=B\int_{0}^{E_{m}}D(E_{x})\int_{0}^{\infty}[f(E)-f(E+eV)]dE_{r}dE_{x}, (2) where the constant B=\frac{4\pi m_{\textrm{e}}e}{h^{3}}, and where m_{\textrm{e}} is the mass of the electron in the electrodes, taken here to equal the free-electron mass, e is the elementary charge, E_{m} is the maximum energy of the tunneling electrons, f(E) and f(E+eV) are the Fermi-Dirac distribution functions in electrodes 1 and 2, respectively, and E_{r}=E_{y}+E_{z} is the transverse part of the total energy E=E_{x}+E_{y}+E_{z} . At zero temperature, Eq. ( 2 ) simplifies to 31 ; 27 \displaystyle J=B\Big[eV\int_{0}^{\mu-eV}D(E_{x})dE_{x}+\int_{\mu-eV}^{\mu}(\mu-E_{x})D(E_{x})dE_{x}\Big], (3) whereas at nonzero temperatures, the transverse energy integral in Eq. ( 2 ) has an analytical form, giving the general solution 31 ; 29 \displaystyle J(V,T)=Bk_{\textrm{B}}T\int_{0}^{E_{m}}D(E_{x}) \displaystyle\times\ln\Bigg\{\frac{1+\exp[(\mu-E_{x})/(k_{\textrm{B}}T)]}{1+\exp[(\mu-E_{x}-eV)/(k_{\textrm{B}}T)]}\Bigg\}dE_{x}, (4) where k_{\textrm{B}} is the Boltzmann constant. III Foundations and main results of the Simmons model Next, we critically review the results of Simmons’ work. This is important as our modeling uses the same foundations, but extends the range of applicability of the original model. To be able to derive an analytical result for a general barrier shape, Simmons makes the key approximation

Read the complete text