Mellin-Space Prony Representability of Linear Viscoelastic Models

This paper establishes a necessary and sufficient condition for the finite Prony representability of linear viscoelastic models by analyzing the alignment of arithmetic pole lattices in the Mellin transform of the complex modulus, thereby providing a complete analytical taxonomy that classifies classical models as finitely representable and fractional or log-normal models as requiring infinite Prony ladders.

The Gist

Imagine you are trying to describe how a piece of silly putty or a thick gel behaves when you stretch it. These materials are viscoelastic: they act like a mix of a spring (which snaps back instantly) and a dashpot (a piston in oil that resists movement slowly).

Scientists have two main ways to describe this behavior:

1. The "Lego" Model (Finite Prony Series): Imagine building a machine out of a specific, countable number of springs and dampers. You have 3 springs and 2 dampers. It's a finite, discrete list. This is easy to understand and build.
2. The "Fog" Model (Continuous/Fractional): Imagine the material isn't made of distinct parts, but is more like a fog or a smooth gradient where relaxation happens at every possible speed simultaneously. This is mathematically complex and "infinite."

The Big Problem:
In the real world, we can only measure materials at specific, discrete speeds (like taking photos at specific intervals). We never see the "fog" directly. So, scientists often try to force the "Fog" models into "Lego" models by approximating them with a huge number of springs. But is this approximation ever exact? Or is the "Fog" fundamentally different from any "Lego" structure, no matter how many pieces you use?

The Paper's Solution: The "Mellin Space" Translator
The author, Dimiter Prodanov, introduces a mathematical tool called the Mellin Transform. Think of this as a special pair of glasses or a translator that converts the messy time-based behavior of the material into a clean, geometric map of poles (points on a graph).

Here is the core discovery, explained simply:

1. The "Lattice" Analogy

Imagine the mathematical map of the material is a city grid.

• The "Lego" Models (Classical): These models have poles that line up perfectly on a grid with integer spacing (like street corners at 1, 2, 3, 4...). They fit together like a perfect puzzle.
• The "Fog" Models (Fractional): These models have poles that are spaced out by weird, non-integer numbers (like 1.5, 2.3, 3.7...). They are like a grid that is slightly tilted or stretched.

2. The Two Rules for "Lego-ness"

The paper proves that a material can be perfectly described by a finite number of springs (a "Lego" model) if and only if two conditions are met:

• Rule A: The Grid Must Align (Lattice Alignment).
  The "tilted" grid of the material must line up exactly with the "integer" grid of the springs. If the material's math has a spacing of 1.5, but your springs only work on a 1.0 grid, they will never match up perfectly. No matter how many springs you add, you'll always have a gap.

  • Result: Models like Cole-Cole or Fractional Zener have "tilted" grids. They are transcendental. You cannot build them with a finite number of springs. You need an infinite "ladder" of springs to approximate them.

• Rule B: The Residues Must Decouple (Residue Compatibility).
  Even if the grids align, the "weights" (residues) of the poles must follow a simple, independent pattern. If the weight of one spring depends in a complicated way on the weight of another, the system breaks.

  • Result: The Cole-Davidson model has a perfect grid (Rule A passes), but the weights are tangled (Rule B fails). So, even though it looks like it should be a Lego model, it's actually a "Fog" model that requires an infinite ladder.

3. The "Infinite Ladder"

For the materials that fail these rules (the "Fog" models), the paper doesn't say "give up." Instead, it offers a new way to build them: The Infinite Prony Ladder.
Imagine a ladder where the rungs get closer and closer together, stretching out to infinity. If you use this infinite ladder, you can perfectly recreate the "Fog" behavior. The paper shows exactly how to construct this ladder using a logarithmic scale (like the Richter scale for earthquakes, where each step is a fixed multiple of the last).

Summary of the "Verdict"

The paper acts as a classifier for all viscoelastic materials:

Material Type	Examples	Can it be built with a finite number of springs?	Why?
Classical	Maxwell, Standard Linear Solid	YES	Their math grids align perfectly with integer steps.
Fractional	Power-law, Cole-Cole, Havriliak-Negami	NO	Their math grids are "tilted" (non-integer spacing). They need an infinite ladder.
Weird Cases	Cole-Davidson	NO	The grid aligns, but the "weights" are too tangled. They need an infinite ladder.
Random	Log-Normal (Gaussian)	NO	The math is a smooth curve with no grid at all. It needs an infinite ladder.

The Takeaway

This paper draws a sharp line in the sand. It tells us that some materials are fundamentally "digital" (finite springs) and others are fundamentally "analog" (infinite continuous spectra).

If you try to force an "analog" material (like a biological tissue or a polymer gel) into a "digital" model with a finite number of parameters, you are mathematically doomed to error, no matter how many parameters you add. You must accept that these materials require an infinite description, or use the specific "infinite ladder" method the author provides to approximate them correctly.

It's like trying to build a perfect circle out of square Lego bricks. You can get close, but you can never get it exact unless you have an infinite number of infinitely small bricks. This paper tells us exactly which materials are circles and which are squares.

Technical Summary

Here is a detailed technical summary of the paper "Mellin-Space Prony Representability of Linear Viscoelastic Models" by Dimiter Prodanov.

1. Problem Statement

Linear viscoelasticity is mathematically described by continuous relaxation functions (distributions of relaxation times). However, experimental data is inherently discrete and band-limited, creating a fundamental gap between continuous theoretical models and empirical observations.

• The Core Question: Given a specific complex modulus $G^*(\omega)$, does it admit an exact finite Prony representation? That is, can the material's behavior be modeled by a finite network of springs and dashpots (a finite sum of exponential terms in the time domain)?
• The Challenge: While classical models (Maxwell, Standard Linear Solid) are known to be finitely representable, fractional models (Power-law, Cole-Cole, Zener) and continuous spectra (Gaussian, Log-normal) are often approximated by infinite sums. Existing numerical methods to recover Prony parameters from frequency data are ill-posed and suffer from instability (Hadamard instability), making it difficult to distinguish between a truly discrete spectrum and a continuous one.
• Goal: To establish a rigorous, analytical criterion to determine whether a given viscoelastic model belongs to the class of finitely representable models or requires an infinite (transcendental) representation.

2. Methodology

The author employs Mellin transform analysis to recast the problem from the time/frequency domain into the complex Mellin domain. This approach transforms the convolution integral of viscoelasticity into a simple multiplicative relationship involving pole lattices.

• Mellin-Space Formulation:
  The constitutive equation in the frequency domain is transformed into the Mellin domain:
  $$\tilde{G}(s) = \tilde{K}(s) \tilde{H}(-s)$$
  Where:

  • $\tilde{G}(s)$ is the Mellin transform of the complex modulus (minus the equilibrium modulus).
  • $\tilde{K}(s) = -\pi e^{i\pi(1-s)/2} \csc(\pi s)$ is the Mellin transform of the spectral kernel, possessing simple poles at all integers $s \in \mathbb{Z}$.
  • $\tilde{H}(s)$ is the Mellin transform of the relaxation spectrum $H(\tau)$.

• The Finite Prony Class ($\mathcal{P}$):
  The paper defines a maximal class $\mathcal{P}$ of relaxation spectra whose Mellin symbols are ratios of Gamma functions (Mellin-Barnes/Fox-H type). A function belongs to $\mathcal{P}$ if its Mellin transform is of the form:
  $$\tilde{H}(s) = H_\Gamma(s) Q(s)$$
  where $H_\Gamma(s)$ is a finite ratio of Gamma functions and $Q(s)$ is an entire function.

• Pole-Lattice Analysis:
  The central insight is that the poles of Gamma factors form arithmetic progressions (lattices). The existence of a finite Prony representation depends on the geometric alignment of these lattices with the integer lattice of the kernel $\tilde{K}(s)$.

3. Key Contributions and Theoretical Framework

A. The Finite Prony Representation Criterion (Theorem 1)

The paper proves that a complex modulus $G^*(\omega)$ admits an exact finite Prony representation if and only if two conditions are met:

1. Lattice-Alignment Condition: The pole lattices of the Gamma factors in $\tilde{G}(s)$ must align exactly with the integer lattice of $\tilde{K}(s)$ (or the reflected lattice of $\tilde{H}$). Specifically, the spacing of the poles ($\Delta_G = 1/|\alpha|$) must be an integer (i.e., $\alpha = 1$). If the spacing is non-integer (fractional), the lattice cannot align with the integers, and finite representability is impossible.
2. Residue-Compatibility Condition: The residues of the poles along the aligned sublattices must satisfy a decoupled first-order recurrence relation. If residues from different pole families are coupled (creating an infinitely overdetermined system), no finite solution exists.

B. Classification of Models

Using this criterion, the paper provides a complete taxonomy of viscoelastic models:

• Finitely Representable (Class $\mathcal{P}$):
  • Maxwell, Kelvin-Voigt, Standard Linear Solid (SLS): These models have rational moduli with integer-spaced pole lattices ($\Delta_G = 1$) and decoupled residues. They correspond to finite mechanical networks.
• Transcendentally Representable (Infinite Prony Ladders):
  • Fractional Models (Cole-Cole, Havriliak-Negami, Fractional Zener): These possess non-integer pole spacings ($\Delta_G = 1/\alpha$ or $\alpha$ where $\alpha \neq 1$). By Lemma 1, these lattices cannot align with the integer lattice, violating the Lattice-Alignment Condition.
  • Cole-Davidson: Although it has integer spacing ($\Delta_G = 1$), it fails the Residue-Compatibility Condition because the residues of its Gamma factors are coupled, preventing a finite solution.
  • Power-Law, Gaussian, and Log-Normal: These have continuous spectra or entire Mellin transforms (e.g., $e^{s^2}$) that are not finite exponential sums. They require infinite Prony ladders.

C. Transcendental Prony Representation (Theorem 2)

For models that are not finitely representable, the paper proves they admit an exact infinite Prony ladder representation. By discretizing the continuous relaxation spectrum using a geometric progression ($\tau_k = \tau_0 q^k$), the discrete sum converges weakly (in the distributional sense) to the continuous spectrum. This provides a constructive method to approximate any viscoelastic model with arbitrary precision using an infinite series.

4. Results

• Analytical Proof of Dichotomy: The paper rigorously proves that the distinction between classical (finite) and fractional (infinite) viscoelasticity is structural, rooted in the arithmetic geometry of the poles in the Mellin domain.
• Rejection of Fractional Models: It is proven that fractional models (with non-integer orders) cannot be represented by any finite network of springs and dashpots, regardless of the number of elements.
• Log-Normal Spectrum: The paper identifies the log-normal distribution (often cited as the "most probable" spectrum via information theory) as fundamentally transcendental, requiring an infinite ladder for exact representation.
• Practical Test: A step-by-step algorithm is provided to test any given modulus:
  1. Compute the Mellin transform.
  2. Extract the pole spacing $\Delta_G$.
  3. Check if $\Delta_G$ is an integer.
  4. If integer, check for residue coupling.
  5. If either fails, the model is not finitely representable.

5. Significance

• Theoretical Unification: The work unifies the treatment of diverse viscoelastic models (power-law, Gaussian, Mittag-Leffler) under a single complex-analytic framework, linking fractional calculus to the arithmetic properties of Gamma function poles.
• Resolution of the Inverse Problem: It resolves the ill-posedness of recovering Prony parameters from data by providing an analytical "yes/no" test for finite representability before attempting numerical fitting.
• Modeling Implications: It clarifies that fractional viscoelasticity is not merely a mathematical convenience but represents a physical reality that cannot be captured by finite mechanical networks. This justifies the use of infinite ladder approximations or fractional operators in modeling soft materials with fractal structures.
• New Paradigm: The paper establishes a new paradigm for classifying materials based on the "Prony representability" of their constitutive laws, distinguishing between materials that are "finitely realizable" and those that are "transcendentally realizable."

In summary, Prodanov's work provides a definitive mathematical boundary between classical and fractional viscoelasticity, demonstrating that the latter requires infinite degrees of freedom (infinite Prony ladders) due to the non-integer arithmetic structure of their singularities in the Mellin domain.