Nonlinear Viscoelasticity

Many materials, from the plastics in our electronics to the tissues in our bodies, possess a fascinating dual nature, behaving as both elastic solids and viscous fluids. This 'viscoelastic' behavior is often described by simple, elegant linear theories where the response is directly proportional to the applied force. However, when these materials are subjected to large deformations or high stresses, this simple picture breaks down, revealing a new world of complex and often counter-intuitive behaviors. This is the realm of nonlinear viscoelasticity, a critical field for understanding how materials truly function under real-world conditions.

This article addresses the fundamental question: what happens when the neat rules of linearity are violated? We will explore the tell-tale signs of nonlinearity and the advanced models required to describe them. The journey will begin by examining the core principles and mechanisms that distinguish nonlinear from linear viscoelasticity. Following this, we will uncover the profound implications of these principles across a wide range of applications and interdisciplinary connections, from the engineering of polymers to the mechanics of living cells.

Imagine you are an engineer building a bridge. You have a new type of plastic you want to use. You test it by hanging a one-kilogram weight from it, and you measure how much it sags over a day. You find it sags by one centimeter. Now, what do you expect to happen if you hang a two-kilogram weight? If your intuition, honed by years of experience with steel, springs, and rubber bands, tells you it should sag by two centimeters, then you have just stumbled upon the cornerstone of a vast and beautiful area of physics: the principle of ​​linearity​​.

For a great many phenomena in our world, the response of a system is directly proportional to the stimulus we apply. Double the force, and you double the deflection. Double the voltage, and you double the current. This elegant property, called ​​homogeneity​​, combined with another property called ​​additivity​​ (the response to two stimuli applied together is the sum of the responses to each applied separately), forms the foundation of linear systems.

In the realm of materials that both flow like a liquid and deform like a solid—viscoelastic materials—this idea is enshrined in the ​​Boltzmann Superposition Principle​​. It tells us something profound: the strain in a material today is a weighted sum of all the stress it has ever experienced in its past, with recent stresses having more influence than those in the distant past. The material has a "memory", but it is a simple, orderly memory. The principle works because it assumes that the rules governing the material’s response don't depend on how much it has already been stretched or squeezed. The material's properties are constant. This is the world of ​​linear viscoelasticity​​, and it is described by beautiful, manageable mathematics.

But nature is often more subtle and mischievous than our simplest models. What happens when we push the material a little harder? We find that this simple, elegant world begins to warp and twist in fascinating ways. These behaviors, which fall under the umbrella of ​​nonlinear viscoelasticity​​, are not mere curiosities; they are dominant in everything from the plastics in our electronics to the living tissues in our bodies.

The first clue that we have left the comfortable world of linearity comes when simple proportionality fails. Let's return to our plastic beam. We perform a very careful set of ​​creep tests​​, where we apply a constant stress and watch the strain (deformation) evolve in time.

In one experiment, we apply a stress of $\sigma = 5 \text{ MPa}$ and find that after 1000 seconds, the strain is $0.0060$. In a second experiment, we double the stress to $10 \text{ MPa}$. Linearity would predict the strain should double to $0.0120$. But instead, we measure $0.0130$. It's a small difference, but it's real. If we triple the initial stress to $15 \text{ MPa}$, we find the strain becomes $0.0220$, much more than the $0.0180$ linearity would suggest.

A powerful way to visualize this is by constructing an ​​isochronous stress-strain diagram​​, where we plot the stress we applied against the strain we measured at the same fixed point in time (say, 1000 seconds). For a linear material, this plot must be a straight line passing through the origin. The slope of this line is the ​​creep compliance​​, $J$, a measure of how much the material deforms under a given stress. Our data, however, plots as a curve that bends upwards. This is a tell-tale sign of nonlinearity: the compliance is not a constant; it depends on the stress itself. The material becomes "softer" or more compliant as the stress increases.

This breakdown of proportionality is a universal signature of the transition from linear to nonlinear behavior. We can also see it in ​​creep and recovery​​ experiments. In the linear regime, a "normalized" creep curve—the strain divided by the applied stress—is a unique property of the material. It doesn't matter if you apply $1 \text{ MPa}$ or $2 \text{ MPa}$; the normalized curve is identical. In the nonlinear regime, this is no longer true. The normalized curve for the $2 \text{ MPa}$ test will be different from the $1 \text{ MPa}$ test, signaling that the rules of the game have changed.

Perhaps the most dramatic and unexpected signs of nonlinearity are not just changes in magnitude, but the appearance of entirely new phenomena. Consider stirring a thick polymer solution. You are applying a shear flow. In the linear world of a simple fluid like honey (a Newtonian fluid), you feel a resistance, and that’s it. But a polymer solution is different.

1. ​​Stress Overshoot:​​ When you suddenly start stirring at a constant rate, the resistance (shear stress) doesn't just rise smoothly to a steady value. It often shoots above its final steady-state value before relaxing back down. The material momentarily resists the change more strongly than it resists the steady flow.
2. ​​Normal Stresses:​​ Even more strangely, the fluid pushes outwards, perpendicular to the direction you are shearing it. Imagine the polymer molecules as long strands of spaghetti. As you stir, these strands align and stretch along the flow direction. Their elastic tendency to recoil creates a tension along the streamlines, which manifests as a pressure pushing outwards. This phenomenon gives rise to the ​​first normal stress difference​​ ($N_1$), a quantity that is zero in the linear regime but is a cardinal feature of nonlinear viscoelasticity. It’s responsible for bizarre effects like the fluid climbing up the shaft of a spinning mixer (the ​​Weissenberg effect​​).

These phenomena—stress-dependent compliance, stress overshoot, and the appearance of normal stresses—are our unambiguous dispatch from the material that we have entered a new, richer, and more complex physical domain.

Another way to explore a material's properties is to wiggle it back and forth and see how it responds. In ​​Dynamic Mechanical Analysis (DMA)​​, we impose a small, sinusoidal strain, $\epsilon(t) = \epsilon_0 \sin(\omega t)$, and measure the resulting stress. In the linear viscoelastic world, the response is also a perfect, phase-shifted sine wave: $\sigma(t) = \sigma_0 \sin(\omega t + \delta)$. The material sings back the same pure note we played. The response is fully described by two numbers: the ​​storage modulus​​ $E'$, representing the elastic part (in-phase with strain), and the ​​loss modulus​​ $E''$, representing the viscous part (in-phase with strain rate).

What happens if we increase the amplitude $\epsilon_0$ of our wiggle, pushing beyond the linear region? The material's response becomes distorted. It's no longer a pure sine wave. Instead, it's a complex, periodic wave containing the original driving frequency $\omega$ plus integer multiples of that frequency: $2\omega, 3\omega, 4\omega$, and so on. These are ​​higher harmonics​​. It’s like plucking a guitar string so hard that you hear a dissonant "buzz" of overtones along with the fundamental note. The simple description of $E'$ and $E''$ is no longer sufficient because it only accounts for the fundamental frequency.

This has given rise to a powerful modern technique called ​​Large-Amplitude Oscillatory Shear (LAOS)​​. In LAOS, we don't ignore the distortion; we measure it. We decompose the distorted stress signal into its constituent frequencies using a mathematical tool called a Fourier series. Now, instead of just two moduli, we have a whole family of them: a storage and loss modulus for the first harmonic ($G'_1, G''_1$), a pair for the third ($G'_3, G''_3$), a pair for the fifth, and so on. This set of ​​nonlinear moduli​​ provides a detailed fingerprint of the material's nonlinear character.

Amazingly, fundamental symmetries of the material dictate what this fingerprint can look like. For most simple, isotropic materials, reversing the direction of shear simply reverses the sign of the stress. A consequence of this simple symmetry is that the stress response to a pure sinusoidal strain can only contain ​​odd harmonics​​ ($\omega, 3\omega, 5\omega, \dots$). The even harmonics are forbidden by symmetry! This is a beautiful example of how deep, underlying principles of symmetry manifest themselves in observable mechanical behavior.

Since the elegant Boltzmann superposition principle is broken, what can we replace it with? We need a new framework.

A part of the problem at large deformations is simply geometry. The linear superposition principle is usually formulated with a "small" or "engineering" strain, like $\varepsilon = \lambda - 1$, where $\lambda$ is the stretch ratio. This is really just the first term in a Taylor series, guaranteed to fail when deformations are not infinitesimal. The very stiffness of a material changes as it deforms. Think of a rubber band: its incremental stiffness (the force needed for the next little bit of stretch) is much higher when it's already highly stretched. A linear model, which assumes a constant stiffness, completely misses this ​​strain-stiffening​​ effect. Any theory of finite-strain viscoelasticity must first account for the underlying ​​nonlinear elasticity​​ of the material.

One clever and widely used approach is ​​Quasi-Linear Viscoelasticity (QLV)​​. Proposed by the great bioengineer Y.C. Fung to describe living tissues, QLV works with a "separation of variables" ansatz. It assumes that the complex nonlinear response can be split into two simpler parts: a purely elastic, nonlinear function that depends only on the current strain, and a time-dependent relaxation function that is independent of the strain magnitude. The superposition integral is then applied not to the strain history, but to the history of the ​​nonlinear elastic stress​​. This allows the model to correctly capture the strain-dependent stiffness while retaining a simplified, linear representation of the material's memory.

For an even more general description, we must confront the fact that stress can affect not just the magnitude of the response but also its timing. The ​​Schapery model​​ provides a powerful and versatile single-integral framework for this. It modifies the linear theory by introducing a small set of stress-dependent functions, $g_0, g_1, g_2$, which scale the instantaneous and transient parts of the response. Most importantly, it introduces a ​​stress-dependent material clock​​. The rate at which the material relaxes and creeps is no longer constant but is scaled by a shift factor $a_{\sigma}(\sigma)$ that depends on the current stress level. Under high stress, the material might "age" or relax much faster than it would at low stress. This leads us to a profound and unifying concept.

For many polymeric materials, there is a magical equivalence between time and temperature. Performing a test at a high temperature for a short time gives the same result as performing it at a low temperature for a very long time. This is the principle of ​​Time-Temperature Superposition (TTS)​​. It allows us to take a series of measurements at different temperatures, each over a manageable time scale (e.g., hours), and then slide them horizontally on a logarithmic time axis to construct a single "master curve" that predicts the material's behavior over decades or even centuries. This works because, for a ​​thermorheologically simple​​ material, temperature affects the rate of all underlying molecular relaxation processes by the same universal factor, $a_T(T)$, beautifully described near the glass transition by the ​​Williams-Landel-Ferry (WLF) equation​​.

This raises a tantalizing question: can stress do the same thing? Can applying a high stress be equivalent to waiting for a longer time? This is the idea of ​​Time-Stress Superposition (TSS)​​, and it is the physical underpinning of the $a_{\sigma}$ term in the Schapery model. In some cases, and over some ranges of stress, this works remarkably well.

However, TSS is intrinsically more limited than TTS. Temperature is a scalar that generally energizes all molecular motions in a uniform way. It's like a conductor telling the entire orchestra to play faster, but keeping the sheet music the same. Stress, on the other hand, is a tensor. High stress doesn't just energize molecules; it can pull them into alignment, change the free volume available for motion, or even break them apart. It can change the relative importance of different relaxation mechanisms, effectively rewriting parts of the "sheet music." This change in the shape of the relaxation spectrum cannot be undone by a simple horizontal shift in time.

Violation of this separability between time and stress effects is a key feature of nonlinear viscoelasticity. It manifests as changes in the shape and curvature of response curves that cannot be reconciled by simple shifting. Understanding when superposition holds—and more importantly, when and why it fails—is the key to decoding the complex but fascinating behavior of the materials that shape our world.

Now that we have grappled with the principles of how materials can be both solid-like and liquid-like, and how this character changes when you push them hard, you might be asking: "So what?" It's a fair question. After all, what is the use of these sprawling equations and distorted loops?

Well, it turns out this "so what" is the difference between paint that flows smoothly from a brush but doesn't drip from the wall and a runny mess. It is the difference between a plastic pipe that lasts fifty years and one that unexpectedly bursts under pressure. And, most remarkably, it is at the heart of how living cells sense their environment and how you, at this very moment, can feel the motion of your own head. The world is profoundly, wonderfully, and sometimes dangerously nonlinear. Let's take a journey and see where these ideas about nonlinear viscoelasticity lead us.

Our first stop is the world of engineering, where controlling how things flow and hold their shape is paramount. Think of any thick liquid you've used: ketchup, shampoo, paint, or a polymer melt destined to become a car part. Their behavior is governed by nonlinear viscoelasticity.

A simple yet profound question we can ask is: when does a material start to behave nonlinearly? Imagine a polymer solution as a tangled mess of long, spaghetti-like molecules. At rest, they're coiled up randomly. If you gently stir the solution, the molecules have time to relax back to their comfortable, coiled state. The viscosity stays constant; this is the linear regime. But what if you stir it very quickly? You are deforming the fluid faster than the molecules can relax. You are pulling the molecular spaghetti straight before it has a chance to recoil. This alignment reduces the resistance to flow, and the viscosity drops. This is called ​​shear thinning​​.

The crossover happens when the time scale of deformation, which is related to the inverse of the shear rate ($1/\dot{\gamma}$), becomes comparable to the material's longest internal relaxation time, $\lambda$. This gives us a beautiful, simple dimensionless number, the Weissenberg number, $Wi = \lambda \dot{\gamma}$. When $Wi \ll 1$, the material has plenty of time to relax, and its response is linear. When $Wi \sim 1$ or greater, nonlinearity kicks in, and phenomena like shear thinning emerge. This simple rule of thumb, born from comparing two time scales, is the key to designing everything from injectable drugs to the processing of plastics.

But sometimes, pushing a fluid harder doesn't just make it flow faster. It can do something much stranger: it can become unstable. Certain nonlinear viscoelastic models predict that as you increase the shear rate, the stress required to maintain the flow can actually go down over a certain range. Nature finds this situation unstable, and the fluid resolves the paradox by spontaneously separating into distinct layers, or bands, one flowing at a low rate and one at a high rate, coexisting side-by-side. This phenomenon, known as ​​shear banding​​, is a dramatic consequence of the nonlinear relationship between stress and strain rate, and our constitutive models can help us predict when it will occur.

The stakes get higher when we move from fluids that flow to solids that are supposed to hold their form. Here, understanding nonlinear viscoelasticity is often a matter of safety and reliability, as it governs how and when materials break.

Consider a plastic component in a machine that vibrates. It's subjected to cyclic loading, over and over again. How long will it last before it suffers from ​​fatigue​​ and fails? A naive guess might be to link the fatigue life to the energy dissipated in each cycle. In the linear world, this dissipation is neatly captured by the loss modulus, $E''$. One might be tempted to measure $E''$ at small strain, calculate the dissipated energy, and predict the material's lifetime.

This is a dangerous trap. The problem is that fatigue failure happens under large strain amplitudes, deep in the nonlinear regime. Here, the simple relationship between dissipated energy and $E''$ breaks down. The stress response becomes distorted, or anharmonic, and the dissipated energy loop is no longer a perfect ellipse. More importantly, the material itself is not static; it is accumulating damage with every cycle. The small-amplitude $E''$ we measured was for the pristine, undamaged material. It tells us nothing about the behavior of the "battle-hardened" material just moments before it fractures. To predict failure, we must embrace the nonlinear, history-dependent nature of the material's response.

Sometimes, failure is more subtle. In many glassy polymers, like the clear plastic in a food container, failure begins not with a crack but with a ​​craze​​. A craze is a microscopic, spongy region filled with tiny voids and highly stretched polymer fibrils. It's a precursor to a crack. What's fascinating is that crazes can form even when the applied load is well below the material's official yield strength. How?

The answer lies in a phenomenon called ​​ratcheting​​. At the tip of a microscopic flaw or a sharp corner in a part, the stress is concentrated. Because the material is viscoplastic—a cousin of viscoelasticity where permanent deformation occurs—each loading cycle can add a tiny, unrecovered bit of strain. This causes the local mean stress to "ratchet up" with every cycle. Crazing is driven primarily by hydrostatic tension (forces pulling the material apart), not just shear. This ratcheting of stress can cause the local hydrostatic tension to build up, cycle by cycle, until it reaches the critical threshold to tear the polymer apart on a microscopic level and nucleate a craze. It's a slow, insidious march toward failure, driven entirely by the material's nonlinear response.

Of course, sometimes materials don't fail with a bang, but with a slow, silent slump. This is ​​creep​​, the tendency of a solid to deform permanently under a constant load. For a plastic shelf holding books, creep is the enemy. Engineers have developed sophisticated nonlinear models, like the Schapery model, to predict this long-term behavior. These models capture the idea that the material's "memory" of its past shape is not constant, but can be influenced by the current stress level. Under high stress, the memory fades faster, and the material deforms more quickly.

How do we take these complex, nonlinear ideas and use them to design an actual car bumper or a medical implant? The answer lies in the fusion of experimental characterization and computational simulation.

On the experimental side, we "listen" to the material's nonlinear dialect. A powerful technique is ​​Large Amplitude Oscillatory Shear (LAOS)​​. We subject a material to a large sinusoidal strain and measure the resulting stress. In the nonlinear regime, the stress response is no longer a perfect sine wave; it becomes a distorted waveform rich with higher harmonics. By using Fourier analysis to break down this distorted signal into its constituent frequencies, we can extract a quantitative "fingerprint" of the nonlinearity. This allows us to see how the apparent stiffness and damping change with amplitude and to validate our constitutive models. And in a moment of mathematical elegance, it turns out that even with all these messy higher harmonics, the total energy dissipated per cycle is still cleanly determined by the first harmonic of the response.

On the computational side, these validated models are implemented in powerful software, often using the Finite Element Method (FEM). This allows engineers to simulate the behavior of an entire object under complex loading conditions. For these simulations, different numerical strategies exist. Some are elegant but are restricted to linear problems. The real workhorses are methods that integrate the material's response step-by-step through time, which can handle the full messiness of nonlinearity, damage, and arbitrary loading histories, albeit at a greater computational cost.

Perhaps the most exciting and rapidly growing application of nonlinear viscoelasticity is not in man-made materials, but in the materials we are made of. Biology, it turns out, is a master of nonlinear soft matter.

Consider a bacterial ​​biofilm​​—the slimy coating that can form on teeth, in pipes, or on medical implants. This is not just a pile of bacteria; it is a complex, structured community encased in a matrix of extracellular polymeric substances (EPS), a gel made of polysaccharides, proteins, and DNA. The mechanical integrity of this matrix is crucial for the biofilm's survival and its resistance to antibiotics. But what part of the matrix provides the strength?

Here, rheology becomes a tool for molecular dissection. We can place an intact biofilm in a rheometer and measure its storage modulus ($G'$), which tells us about its solid-like stiffness, and its loss modulus ($G''$), which tells us about its liquid-like dissipation. Then, we can perfuse the sample with a specific enzyme, say a protease that chews up only proteins. By watching how $G'$ and $G''$ change in real time as the protein network is degraded, we can directly quantify the load-bearing role of those specific fibers. It is like selectively removing the steel rebar from a concrete structure to see how much it weakens. This approach gives us a profound understanding of the structure-function relationship in these complex living materials.

To end our journey, let us consider one of the most elegant examples of functional nonlinear mechanics in all of biology: the hair cells in your inner ear that constitute your sense of balance. When you rotate your head, inertia causes the fluid in your semicircular canals to lag, deflecting a gelatinous structure called the cupula, which in turn deflects a tiny bundle of "hairs" (stereocilia) on top of a hair cell. This deflection pulls on minuscule molecular filaments called tip links, which act like "gating springs" that open and close ion channels.

The probability of a channel being open or closed is governed by the energy difference between the two states, a classic problem in statistical mechanics that follows a Boltzmann distribution. The inevitable consequence of this simple physics is that the channel's open probability—and thus the electrical current flowing into the cell—is a ​​sigmoidal​​ (S-shaped) function of the hair bundle's displacement. It's linear for small displacements but saturates for large ones; you can't open more than all the channels, and you can't close more than all of them. This is the first nonlinearity.

The second comes from the neuron connected to the hair cell. It has a resting firing rate, and while it can fire faster, it cannot fire at a negative rate. Its output is clipped at zero. The combination of sigmoidal compression and inhibitory clipping means that the signal sent to your brain is inherently asymmetric and nonlinear for large, fast head movements. It is a beautiful illustration of how fundamental physical constraints on a material system (nanoscale biological machinery) directly shape a macroscopic biological function—your very perception of the world.

From industrial fluids and failing plastics to biofilms and the exquisite sensors in our own bodies, the ideas of nonlinear viscoelasticity provide a powerful and unifying language. It is a science that reveals the rich and complex character of the materials that constitute our modern world, both built and living.