The $i\varepsilon$-Prescription for String Amplitudes and Regularized Modular Integrals

This paper establishes that the analytic continuation of one-loop string amplitudes via a string-theoretic $i\varepsilon$-prescription is equivalent to a regularization using generalized exponential integrals, yielding exact expressions for the imaginary and real parts of various open and closed string amplitudes that agree with the Hardy-Ramanujan-Rademacher circle method approach.

The Gist

Imagine you are trying to calculate the total "energy" or "cost" of a complex journey taken by a tiny string in the universe. In the world of string theory, these calculations often involve summing up an infinite number of possibilities. However, when physicists try to do the math, they often run into a wall: the numbers blow up to infinity. It's like trying to add up a list of numbers where the last few entries are infinite; the total becomes meaningless.

This paper, written by Jan Manschot and Zhi-Zhen Wang, tackles a specific problem: How do we fix these "infinite" calculations to get a real, usable answer?

Here is a breakdown of their approach using simple analogies:

1. The Problem: The "Infinite" Roadblock

In physics, there is a standard trick called the $i\epsilon$-prescription (think of it as a "safety valve" or a "detour sign"). In standard particle physics (Quantum Field Theory), this trick helps avoid infinite results by slightly shifting the path of calculation into a different dimension (imaginary numbers) just long enough to bypass the singularity, then shifting it back.

The authors ask: Does this same trick work for strings?
Strings are more complex than particles; they are like tiny loops or ribbons. Their "journey" isn't just a line; it's a surface (like a donut shape called a torus). When these surfaces stretch out too long, the math breaks down. The authors wanted to prove that the string-theory version of this "safety valve" works and gives the same result as other known methods.

2. The Solution: Two Different Maps to the Same Treasure

The paper compares two different ways to navigate this mathematical minefield:

• Method A: The "Wick Rotation" Detour (The $i\epsilon$-Prescription)
  Imagine you are driving a car on a road that suddenly turns into a bottomless pit. The $i\epsilon$-prescription is like saying, "Okay, instead of driving straight into the pit, let's briefly drive on a parallel road in a parallel universe (the complex plane) to get around the hole, and then come back to our road."

  • The Paper's Claim: They show that if you take this detour for string amplitudes, the math works out perfectly. The "imaginary" part of the journey (the detour) actually tells us something physical: it represents the decay rate of the string (how fast it falls apart).

• Method B: The "Mathematical Filter" (Regularized Modular Integrals)
  This is an older, more abstract method used by mathematicians. Instead of driving around the hole, you use a special filter (called Generalized Exponential Integrals) to subtract the infinite parts before you even start adding them up. It's like using a sieve to remove the sand before weighing the gold.

3. The Big Discovery: The Maps Match

The authors proved that Method A and Method B give the exact same answer.
They showed that taking the "detour" (Method A) is mathematically identical to using the "filter" (Method B). This is a huge deal because:

• It confirms that the string-theory "safety valve" is valid.
• It allows physicists to use the "filter" method to get exact formulas for the imaginary part of the answer (the decay rate) without having to do the messy detour every time.

4. The "Temperature" Analogy

One of the most interesting findings involves Open Strings (strings with ends, like a rubber band).
When calculating the energy of these strings, the authors found the answer looks like a recipe that mixes three different "temperatures" together.

• Imagine you have a pot of soup. The final taste depends on the temperature of the water, the temperature of the stove, and the temperature of the room.
• In their math, the final answer is a combination of three "partition functions" (which are like thermometers measuring the state of the string) at different temperatures.
• The Magic: Even though the individual temperatures change depending on how you set up your calculation (a variable they call $T_0$), the final sum of the three temperatures is always the same. The universe doesn't care how you set the thermostat; the total energy is constant.

5. The "Circle Method" vs. The "Exponential Method"

The paper also compares their new "filter" method with a famous technique from number theory called the Hardy-Ramanujan-Rademacher Circle Method.

• The Circle Method: Think of this as counting the number of ways to arrange coins in a circle. It uses complex patterns (Ford circles) to sum up the answer. It's very precise but can be slow to compute.
• The Exponential Method: This is the authors' new "filter" approach. It's like using a calculator that handles the infinite parts automatically.
• The Verdict: They proved these two very different mathematical languages describe the same reality. The "Exponential Method" is often faster for computers to calculate, while the "Circle Method" gives a beautiful, deep connection to number theory.

Summary of What They Actually Did

• Proved Equivalence: They showed that the "detour" method ($i\epsilon$) and the "filter" method (Regularization) are mathematically identical for string amplitudes.
• Found Exact Formulas: They derived exact formulas for the "decay rate" (imaginary part) of strings, which can be written down clearly without needing a computer.
• Applied to Real Cases: They tested their formulas on specific types of strings (Type I superstrings) and showed they match previous high-precision calculations.
• Numerical Efficiency: They showed that their new "filter" formulas are often faster for computers to calculate than the traditional "Circle Method," especially when high precision is needed.

What they did NOT do:
They did not apply this to clinical uses, black hole physics directly, or new particle accelerators. They stayed strictly within the realm of calculating the mathematical values of string theory amplitudes to ensure the theory is consistent and finite. They also did not solve the "double-copy" problem (relating open and closed strings) completely, but they laid the groundwork for it.

In short, the paper is a mathematical bridge. It connects two different ways of fixing broken string calculations and proves they lead to the same destination, giving physicists a more reliable and faster toolkit for understanding the vibrations of the universe's fundamental strings.

Technical Summary

Problem Statement
One-loop string amplitudes, particularly those involving closed and open strings, often manifest as integrals over the moduli space of Riemann surfaces (specifically the fundamental domain of the torus for closed strings and related geometries for open strings). A central challenge arises when these integrals are infrared (IR) divergent. This divergence occurs when the integrand, expanded as a $q$-series, contains negative powers of $q$ (associated with tachyonic modes or massless states), leading to non-convergence as the imaginary part of the modular parameter $\tau$ approaches infinity.

While standard regularization techniques exist for integrals where only one index of the $q$-expansion is negative, more severe divergences occur where both indices are negative. Furthermore, there is a need to rigorously define the analytic continuation of these amplitudes to ensure unitarity and to extract physical quantities such as mass shifts and decay widths (imaginary parts). Two distinct approaches have been developed to address these divergences:

1. The $i\epsilon$-prescription, a string-theoretic analog of the Feynman prescription in quantum field theory, which involves Wick-rotating the proper time of long tubes in the worldsheet to Lorentzian signature.
2. Regularized Modular Integrals, motivated by analytic number theory and topological quantum field theory, which utilize generalized exponential integrals to regularize the divergent terms.

Although numerical evidence suggested these two methods yield identical results for the bosonic string vacuum amplitude, a rigorous proof of their equivalence and a general framework for applying them to various amplitudes (including open strings and superstring sectors) were lacking.

Methodology
The authors employ a dual approach to evaluate and compare these regularization strategies:

1. Contour Deformation and Complexification:

   • For closed strings, the integration domain (the fundamental domain $\mathcal{F}$) is complexified. The $i\epsilon$-prescription is implemented by deforming the integration contour in the complexified Teichmüller space. Specifically, the proper time $t_E$ is extended to a complex variable $t$, and the contour is rotated from the Euclidean real axis to the Lorentzian imaginary axis at a cutoff $T_0$.
   • For open strings, the integration contours (annulus and Möbius strip) are modified to avoid singularities at the cusps ($\tau=0$ and $\tau=1/2$). The authors introduce contours $\Gamma(T_0)$ consisting of vertical lines and semi-circles of radius proportional to $1/T_0$, effectively replacing the divergent regions with finite arcs.

2. Analytic Evaluation via Special Functions:

   • The authors expand the integrands (modular forms) into their Fourier series (involving degeneracies $F(n)$ at each mass level).
   • The integrals over the deformed contours are evaluated term-by-term. This leads to expressions involving generalized exponential integrals $E_s(z)$, which naturally arise from integrating terms of the form $y^{-s}e^{-4\pi my}$.
   • The authors derive exact closed-form expressions for the imaginary parts of the amplitudes, which correspond to physical decay widths, and provide convergent series for the real parts.

3. Comparison with the Circle Method:

   • The results derived via the $i\epsilon$-prescription and exponential integrals are compared against the Hardy-Ramanujan-Rademacher Circle Method (as applied by Eberhardt and Mizera).
   • The Circle Method evaluates the same amplitudes by deforming the contour to an infinite set of Ford circles, resulting in expressions involving arithmetic exponential sums (reminiscent of Ramanujan or Kloosterman sums).
   • The authors prove the equivalence of the two methods using the residue theorem and contour deformation arguments, demonstrating that the integral over the $i\epsilon$-deformed contour is identical to the integral over the Ford circle contour.

Key Contributions and Results

• Proof of Equivalence: The paper rigorously proves that the analytic continuation based on the $i\epsilon$-prescription is mathematically equivalent to the regularization using generalized exponential integrals for both closed and open string one-loop amplitudes. This is demonstrated by showing that the contour deformation required to move from the $i\epsilon$-contour to the fundamental domain (or vice versa) yields identical values, provided the integration cycle is chosen correctly.
• Exact Expressions for Amplitudes:
  • Closed Strings: The authors provide exact expressions for the zero-point (vacuum) and two-point amplitudes of bosonic closed strings. The imaginary part is derived in closed form (e.g., Eq. 3.44), while the real part is expressed as a sum over mass levels involving exponential integrals.
  • Open Strings: A general formula (Eq. 4.72) is derived for one-loop open string amplitudes with weakly holomorphic modular form integrands. This formula expresses the amplitude as a linear combination of partition functions at different "temperatures" (dependent on the cutoff $T_0$), yet the total sum is independent of $T_0$.
  • Type I Superstring: The methods are applied to the Ramond-Ramond (RR) sector vacuum amplitude and planar/non-planar two-point amplitudes of Type I superstring theory. New expressions for these amplitudes are derived using both the exponential integral method and the Circle Method.
• Numerical Convergence Analysis: The paper analyzes the numerical performance of both regularization strategies.
  • The Modular Regularization Method (using exponential integrals) is shown to have exponential convergence rates with respect to the truncation of the mass level sum, provided the parameter $T_0$ is chosen optimally.
  • The Circle Method (using arithmetic sums) exhibits polynomial convergence rates for high accuracy.
  • The authors provide a comparison (Table 6) highlighting that while the Circle Method is powerful for microstate counting, the exponential integral method offers superior numerical efficiency for evaluating specific amplitude values and provides manifest closed forms for the imaginary parts.

Significance
The paper establishes a unified framework for handling divergent one-loop string amplitudes, bridging the gap between string-theoretic physical prescriptions ($i\epsilon$) and mathematical regularization techniques (modular integrals/Circle Method).

• Physical Interpretation: By providing exact closed-form expressions for the imaginary parts of the amplitudes, the work clarifies the calculation of decay widths and mass shifts for string states, which are crucial for understanding the stability and unitarity of the theory.
• Methodological Robustness: The demonstration that the $i\epsilon$-prescription and modular regularization yield identical results validates the use of the $i\epsilon$-prescription as a physically motivated and mathematically sound tool for string perturbation theory.
• Computational Utility: The derived formulas (specifically Eq. 4.72) offer a practical and computationally efficient alternative to the Circle Method for evaluating one-loop amplitudes, particularly for cases where high numerical precision is required.
• Generalizability: The approach is formulated to handle generic weakly holomorphic modular forms, suggesting applicability to higher-point functions and other sectors of string theory, although the paper focuses on zero- and two-point functions as primary examples.

The authors conclude that while the two strategies (Modular Regularization vs. Circle Method) produce superficially different expressions, they are fundamentally equivalent, with the choice of method depending on the specific requirements for convergence speed and the need for closed-form analytic results.