Sunday, November 11, 2007

F#, Regge Calculus, and other interludes

Still working on Causal Dynamical Triangulations, I've also taken some time out to (start to) learn a delightful new functional programming language, F#, with the help of a good site, online journal, and this book. Mixing functional, imperative, and object-oriented programming with good mathematics features and the .NET framework allows me to blend work and academics. The goal is to write a first-cut CDT program in F# and refine as necessary ....

CDTs depend upon Regge calculus, which is essentially a prescribed way for dividing up spacetime into a discrete lattice (simplexes) while adhering to the Einstein field equations. (Regge calculus is explained in detail in Chapter 41 of Gravitation, aka the Big Black Book.)

In particular, you divide up a smooth 4-manifold into a collection of 4-d simplexes, in the same way you can divide up a 3-sphere into the (2-)triangles of the icosohedron. If you pressed the icosohedron perfectly flat onto a plane, you would see deficit angles or hinges that essentially reflect the its spherical curvature. In the same way, you analyze the simplices of the 4-d manifold to determine its curvature.

If you are extraordinarily careful in how you setup your lattice, you can perfectly constrain your volume using just fixed edge lengths. (For example, a tiling of triangles constrains a surface rigidly, because the tiling cannot be deformed without changing edge lengths. A tiling of squares does not, because the squares can be squashed sideways into parallelograms without changing their edge lengths.)

And if you do this, you can now examine that volume (or brane, or bulk ) by solving the Einstein equations expressed in terms of conditions on the hinge angles (which are themselves functions only of the edge lengths). This is exciting because you can now program a computer (carefully) to solve problems that don't lend themselves to analytic solution, which allows you to do interesting things:

Discrete quantum gravity in the framework of Regge calculus formalism

In the past few weeks, I've also taken time to read up on a few areas of interest.

Perhaps one of the more interesting recent occurrences is the recent re-examination of the 2nd law of thermodynamics, long thought to be inviolable and one of the most solid foundations of physics by such luminaries as Maxwell, Einstein, Eddington, and Brilloun.

Physics is always fun, check your assumptions at the door!

First, a nice publicly accessible summary:

Why Do We Believe in the Second Law?, T. Duncan

The Foundations of Physics journal, Volume 37, Number 12/December 2007 is devoted to this extraordinary topic (unfortunately, e-journal subscriptions required to view), starting with Geraard t'Hooft's editorial, which gives a broad summary of the scope of the papers in the journal.

Next, we have:

The Second Law of Thermodynamics: Foundations and Status
, by D.P. Sheehan

This paper gives a broad overview on three classes of discussion regarding the Second Law now underway: ideal gases, quantum perspectives, and interpretations.

Information Loss as a Foundational Principle for the Second Law of Thermodynamics, by T.L. Duncan and J.S. Semura

This paper explores in detail the concept of information loss as being the fundamental explanation for entropy, essentially casting the 2nd law from "Entropy always increases for irreversible processes" to "Information is always lost for irreversible processes". The authors further argue that all classical derivations of the 2nd law using "entropy" actually incorporate, explicitly or implicitly, information loss as the mechanism.

As an example, Maxwell's demon is shown to be able to violate the 2nd law on the basis of having information -- in particular, knowing how to sort fast-moving from slow-moving particles. However, creation of that information eventually involves the deletion of a bit of information from storage -- for a subsequent Kln2 energy cost -- which is argued as being the source of entropy. (All faults are mine, not the authors, if I've paraphrased these arguments incorrectly.)

Jean E. Burns, in Vacuum Radiation, Entropy, and Molecular Chaos, makes a very interesting extension to classical entropy models for isolated systems. Classical thermodynamic models separate the system from the environment. The canonical example is the refrigerator, which decreases temperature (thus entropy) locally, but at the expense of expelling even more heat (thus increasing entropy) in the environment. The entropy/heat loss inside the system is outweighed by the entropy/heat gain in the environment.

However, what happens for an arbitrarily large system (such as the universe), where there is no external reservoir? Burns argues that vacuum radiation provides the mechanism for entropy increase.

Perhaps of most interest to biologists and science-fiction fans is Sheehan's Thermosynthetic Life, which postulates the existence of life forms deriving their energy solely from thermal energy. In addition to searches for extremophile lifeforms (such as bacteria near volcanic vents) that fit this profile, it provides an engaging test into the 2nd Law, because such life-forms may well violate it!

And, continuing on with our examinations into entropy and information theory, we have:

Information Recovery from Black Holes, V. Balasubramanian, D. Marolf, and M. Rozali

This first-place prizewinning essay of the 2006 Essay Competition of the Gravity Research Foundation provides insights into two crucial questions:
  1. Why do classical black holes have finite entropy equal to a quarter of the horizon area?
  2. How does information escape from an evaporating black hole?
In essence (modulo a great many technical arguments), the answer to both of these questions is that the finite mass black hole, representing a finite number of energy states N, therefore possesses a discrete energy spectrum. In general, discrete spectra are quantum-mechanically non-degenerate, so knowledge of the precise energy and other (commuting) conserved charges determines the quantum state. But General Relativity charges are generically given by boundary terms; thus, the entire state of the black hole resides in the boundary (asymptotic region), available to all observers.

This is at odds with classical GR, because there exist unobservable regions within the black hole that are causally separated. However, in the quantum mechanical case the Heisenberg Uncertainty principle dictates that a "Heisenberg recurrence time" exists. This can be thought of as a sort of spontaneous large thermal fluctuation in which the black hole may be replaced by a ball of expanding hot gas. Although the gas will re-collapse to form another black hole on a relatively short time scale, during the span of its existence the full details of the black hole's internal state are visible from infinity.

And thus we see that at the quantum level, the event horizon becomes ill-defined, and quantum mechanics, entropy, and information theory collaborate against General Relativity to allow what was previously thought to be impossible.

And in a final bit of fun, we see a method for efficiently converting black holes into gravitational waves:

Black Hole Bremsstrahlung: Can it be an efficient source of gravitational waves?

Taking a 2 solar mass black hole traveling at .38c and converting 90% of its rest mass into a lobe-shaped pulse with width deltaU ~ 40 for 10E40 GeV^2 sounds exciting!

No comments: