Monthly Archives: March 2015

Transparent Metals 2

Recently I’ve been reading more about anharmonic properties of crystals, for instance metals. There are phenomena, such as thermal expansion, or electrical conductivity (resistivity), that appear to primarily be related to imperfections in the crystal structure, or to non-harmonic behaviour of the material. A material with a purely harmonic response should have no resistance, and should have no dimensional change when heated or cooled. Instead of thinking of the flow of electrons through a material, one might think of a wave function for the distributed gas of electrons.

This is somewhat prompted by a reading of Blakemore’s book on solid state, mentioned in my post about a year ago (April 20th, 2014) on the topic of transparent metals. I had occasion to return to Blakemore’s book again because of a new investigation. And he is as stimulating as ever!

So here’s the basic idea: Instead of thinking about the flow of electrons (as particles), think about propagation of a wave. That leads one to thinking about optics, and indexes of refraction, and the possibility of focusing etc. For instance, it is standard to talk about the speed with which electrons flow through a metal, and compare different models in terms of the wildly different speeds they predict. Do the three noble metals — copper, silver and gold — have the same speeds of electron flow? Otherwise said, would they have the same index of refraction? Can we make a copper-silver-gold alloy that can act as a lens or a waveguide?

Conversely, one can think of traditional wave models, and consider them in terms of particle flow models, eg the resistance (or conductivity) of glass.

Just some late-night idle musing. Maybe something will intrigue you.

Best wishes,
Ken Roberts
14-Mar-2015

Belgacem Paper — Lambert W Function

A recent paper by C. H. Belgacem deserves mention, as a good illustration of how to use the Lambert W function to explicitly solve a problem in semiconductor design.  Previous methods required iterative solution of a design equation.  That can be time-consuming and also has the disadvantage that there is no analytic formula for the solution of the design equation.

The paper is “Explicit Solution for Critical Thickness of Semicircular Misfit Dislocation Loops in Strained Semiconductor Heterostructures”, by Chokri Hadj Belgacem, published online 01-March-2015 in the journal “Silicon”.

I am particularly appreciative of Belgacem’s paper because of his reference 15, to a paper by Willams in 2005.  Williams paper was not previously familiar to me.  It has some really stimulating ideas.  I cannot usefully contribute to the discussions in Belgacem’s field of semiconductor design, but acknowledge his help in finding the discussions around the Williams paper.  That relates to the task of finding exact solutions of the Schrodinger equation.

Best wishes,
Ken Roberts
11-Mar-2015

Lane-Emden Equation

The Lane-Emden equation is a second order differential equation which is used to model a spherically symmetric gas cloud, eg a stellar interior. Although it is a considerably simplified model (eg, no rotation), it still provides a good starting point. Traditionally the LE equation is written in terms of a linear independent variable, say x, which represents the relative radius of the gas cloud. The variable x goes from 0 (center of cloud) to 1 (outer boundary of cloud). With the assumption of a particular type of model of the gas cloud’s thermodynamics — called a polytropic model — one ends up with the LE equation. The model is characterized by a parameter n, called the polytropic index (n is a non-negative real number, not just an integer) which determines a function f (which depends upon the choice of polytropic index n), which in turn determines the density profile of the cloud, and thereby its other profiles, eg its mass profile.

It turns out, however, that the function f is an even function, and its power series expansion, for instance, involves only even powers of x. For instance, f(x) equals 1 – (1/6)*x^2 + (n/120)*x^4 + … and so on, with terms for x^6, x^8 etc. So one asks, whether it might be useful to write the LE equation (a second order differential equation) in terms not of independent variable x, but in terms of independent variable s = x^2. The variable s is the relative surface area of a spherical shell of relative radius x. At x=1 (outer boundary of the gas cloud), the value of s=1, ie relative area of that outer shell, is also 1.

There is a possibility for getting some physical insight from such a rewrite of the LE equation. The physics which is being modeled involves transport of energy and force between nested shells, and such transport may be conceptually more meaningful if it considered as a function of relative area rather than relative radius.

I have prepared a short pdf (2 pages) which describes such a rewrite of the Lane-Emden equation. Nothing new there; I’m sure such a rewrite has been done by others, as the LE equation has been a subject of study for over a century, and there is a considerable literature. Still, it may be of interest to others who are working their way into the details of the Lane-Emden equation and the stellar interior models for which it is a starting place. I give a few references, including to Chandrasekhar’s classic book on Stellar Interiors, which is still a better explanation than some of the more recent books, as it illustrates some motivations. And also a reference to a very interesting little paper by Klaus Rohe, who used Python to calculate rational expressions for the first 15 coefficients of a power series representation of the LE equation (ie, up to the x^28 term), as expressions in the polytropic index n. His paper is at Arxiv 1409.2008 if you want to go there directly.

Best wishes,
Ken Roberts
07-Mar-2015

Link to pdf file mentioned above, with my rewrite of Lane-Emden equation using relative area…
https://lasi2.wordpress.com/wp-content/uploads/2015/03/lane-emden-rewrite.pdf