26. Eye on ArXiv: 24 Apr 2008 – NSR superstring measures

Today I decided to split my review of new papers in archives which were of interest to me into separate posts. This will actually force me to read each paper more carefully and hopefully will also increase the quality of the review I will start with reviewing the paper by

A.Morozov, “NSR Superstring Measures Revisited”

It is highly technical, definitely not for general audience – probably because the problem discussed is extremely deep and lies in the foundatations of string theory.

Its essence is the following. Let us start with bosonic string theory. To study it in the first approximation, we work with the corresponding partition function. It does not carry much information about the theory as it is (with zero source terms in the effective action). For example, calculating this partition function, we can find that the critical dimension of the bosonic string is ; for other effective action (after integrating all other degrees of freedom out) contains  the Liouville mode (it corresponds to longitudinal fluctuations of the string – the world sheet of the latter wants to get wrinckled), in the corresponding part of the effective action disappears.

This is pretty much all we learn from the partition function alone. To get more information, we want to introduce source terms in the effective action and calculate correlation functions of the string degrees of freedom. Tree level calculations are rather trivial, but at the higher loop levels (say, we want to understand how string effective action gets renormalized) technicalities become more and more serious.

The reason is that, as you may know, loop Feynman diagrams in string theory correspond to world sheets with non-trivial topology and one needs to sum over them. For two loop calculations, corresponding world sheets are Riemann surfaces with genus 2, etc.

Long time ago (in 1986) Knizhnik and Belavin explained how to sum over such surfaces: summation over p-loops reduces to the integration over the space of complex structures of genus Riemann surfaces with a certain measure.

The form of this measure is determined as follows: if we multiply it on where is the matrix of surface periods, then the resulting function will be non-zero everywhere on and holomorphic. This bare fact will fix the form of the measure for . The measure is basically the Mumford measure on the moduli space of Riemann manifolds.

Suppose now we are interested in physics of NSR superstrings (in critical dimension ) and again want to calculate loop corrections to stringy amplitudes. We again clearly need to somehow integrate over Riemann manifolds but the measure should be modified because we now have supermoduli. The corresponding measure for two-, three- and four-loop amplitudes was found relatively early (again, 1986-87), but after that the progress got stuck. The reason was that naive integrating supermoduli out to find analogue of the Mumford measure did not work and proper calculations were really hard to do (they were finally elaborated by D’Hoker and Fong in 2001, 14 years later!, and only for genus 2 so far).

Instead of direct calculation, people were trying a lot to make an educated guess concerning the form of the measure for , and this actually led to noticeable results only recently (this year, to be precise )

The paper by Alexey Morozov is a very serious review of such attempts and basically contains all we know about the NSR measure today.