aiwiki.page
English
Science / Physics / non-newtonian-fluid

Non-Newtonian Fluid

A non-Newtonian fluid has a flow response that cannot be described by a constant viscosity independent of deformation rate and history.

37 keywords10 linked from8 not yet writtenWritten by AI
RheologyViscosityInternational Sy…WaterTemperaturePressureTurbulenceShear ThinningNon-Newton…

A non-Newtonian fluid is a material whose flow behavior cannot be fully described by Newton’s law of viscosity: the proportional relationship between shear stress and shear rate, with a viscosity that is constant at a specified temperature, pressure, and composition. Its resistance to deformation may instead depend on how rapidly it is deformed, how long deformation has continued, or its previous mechanical history. Some non-Newtonian fluids also store elastic energy and develop stresses absent from an ideal Newtonian liquid.

Non-Newtonian behavior occurs in everyday materials such as toothpaste, paint, yogurt, shampoo, and concentrated cornstarch suspensions, as well as in polymer melts, drilling fluids, fresh concrete, and blood. It matters wherever materials must be pumped, mixed, spread, extruded, swallowed, or transported through narrow passages. The subject belongs to rheology, the study of deformation and flow, and connects microscopic structure with macroscopic motion.

“Non-Newtonian” describes a broad family of responses rather than one mechanism or substance class. A material may exhibit several responses simultaneously, and its classification can change with the experimental timescale or range of applied stresses. Understanding these fluids therefore requires both carefully defined measurements and models that distinguish rate dependence, yielding, structural evolution, and elastic memory.

Mechanical definition and constitutive description

Consider a liquid confined between two parallel plates. If one plate moves while the other remains stationary, adjacent layers of liquid slide past one another. For an idealized velocity field vx(y)v_x(y), the shear rate is

γ˙=∂vx∂y.\dot{\gamma}=\frac{\partial v_x}{\partial y}.

Shear stress τ\tau is the tangential force per unit area transmitted between those layers. For a Newtonian liquid in simple shear,

τ=ηγ˙,\tau=\eta\dot{\gamma},

where η\eta is the dynamic viscosity. In the International System of Units, stress is measured in pascals, shear rate in inverse seconds, and dynamic viscosity in pascal-seconds. At approximately 20 ∘C20\,^\circ\mathrm{C}, pure water has a viscosity close to 1.0×10−3 Pa s1.0\times10^{-3}\ \mathrm{Pa\,s}, illustrating the scale of a low-viscosity Newtonian liquid.

The qualification “at specified conditions” is essential. A Newtonian fluid’s viscosity can vary substantially with temperature, pressure, or composition without becoming non-Newtonian. What distinguishes non-Newtonian behavior is the additional dependence on deformation rate, deformation history, or other aspects of the flow. Nor does an irregular-looking motion necessarily indicate non-Newtonian behavior: a Newtonian liquid can develop turbulence, while a strongly non-Newtonian material can flow smoothly.

For steady simple shear, an apparent viscosity can be defined as

ηapp=τγ˙.\eta_{\mathrm{app}}=\frac{\tau}{\dot{\gamma}}.

This quantity is useful but is not necessarily a material constant. Its value may depend on the imposed shear rate, the duration of the measurement, and the sample’s preparation. Reporting one viscosity number without those conditions can therefore conceal important differences between materials.

A complete description generally requires a relation between the stress tensor and the deformation of the material. For an incompressible Newtonian fluid,

σ=−pI+2ηD,\boldsymbol{\sigma}=-p\mathbf{I}+2\eta\mathbf{D},

where pp is pressure, I\mathbf{I} is the identity tensor, and D\mathbf{D} is the symmetric part of the velocity-gradient tensor. A non-Newtonian constitutive equation replaces or extends this simple relationship. It may prescribe a rate-dependent viscosity, include internal structural variables, or evolve an additional elastic stress.

The distinction is especially important outside simple shear. Two fluids with nearly identical shear-viscosity curves may behave very differently when stretched, squeezed, or driven around a curved channel. Shear viscosity alone does not determine all the stresses generated by a complex flow, and no single scalar measurement can fully characterize every non-Newtonian material.

Shear thinning and shear thickening

The most familiar form of non-Newtonian behavior is a steady viscosity that changes with shear rate. In shear thinning, apparent viscosity decreases as shear rate increases. Polymer solutions, many paints, and numerous food dispersions exhibit this response over substantial operating ranges. A sauce can consequently resist slow flow on a plate yet move readily when stirred or forced through a nozzle.

A common empirical approximation is the power-law model,

τ=K∣γ˙∣n−1γ˙,ηapp=K∣γ˙∣n−1,\tau=K|\dot{\gamma}|^{n-1}\dot{\gamma}, \qquad \eta_{\mathrm{app}}=K|\dot{\gamma}|^{n-1},

where KK is the consistency index and nn the flow index. Values n<1n<1 describe shear thinning, n=1n=1 gives Newtonian behavior, and n>1n>1 describes shear thickening. The units of KK are Pa sn\mathrm{Pa\,s^n}, so consistency indices from fluids with different flow indices cannot be compared as though they were ordinary viscosities. The model is usually a local approximation: extrapolating a shear-thinning power law to zero shear rate predicts an infinite viscosity, even when the actual material has a finite low-rate plateau.

In many polymer liquids, shear thinning reflects the competition between deformation and molecular relaxation. At low rates, molecular configurations remain close to their equilibrium distribution. At higher rates, chains become oriented and their conformations and constraints change, reducing resistance relative to the imposed rate. In structured dispersions, thinning may instead result from the disruption of particle aggregates or alignment of elongated objects. Similar flow curves can therefore arise from different microscopic processes.

Shear thickening is an increase in apparent viscosity with shear rate or stress. It is especially important in dense particle suspensions. Thickening can be gradual, or it can occur abruptly over a narrow stress interval. Concentrated cornstarch in water is a familiar example, although its response depends strongly on concentration, preparation, and confinement; not every cornstarch mixture displays the same effect.

For many dense suspensions, a major mechanism involves a stress-induced transition from predominantly lubricated particle interactions to more numerous frictional contacts. At sufficiently high concentration, the resulting contact network can sharply increase flow resistance or produce jamming. The tendency of a sheared packing to expand, called dilatancy, can reinforce this resistance when expansion is constrained. Shear thickening and dilatancy are related but are not interchangeable concepts. Impact experiments also introduce transient effects, including rapidly growing jammed regions and interactions with boundaries, that cannot be inferred from a steady flow curve alone.

Yield stress and the boundary between liquidlike and solidlike behavior

Some materials resist small applied stresses but flow continuously once the stress exceeds a characteristic level. This behavior is described by a yield stress. Toothpaste, concentrated emulsions, mineral pastes, and many drilling fluids can remain approximately at rest under their own weight while flowing when squeezed or pumped. Their ability to support stress reflects an internal network, crowded arrangement, or other mechanically connected structure rather than simply a large Newtonian viscosity.

The Bingham model represents yielding through a threshold τy\tau_y followed by a linear relation between additional stress and shear rate. In ideal simple shear,

γ˙=0when∣τ∣≤τy,\dot{\gamma}=0 \quad \text{when}\quad |\tau|\leq\tau_y,

and, in a yielded region,

τ=τysgn⁡(γ˙)+ηpγ˙,\tau=\tau_y\operatorname{sgn}(\dot{\gamma}) +\eta_p\dot{\gamma},

where ηp\eta_p is the plastic viscosity. A more flexible description, the Herschel–Bulkley model, replaces the linear post-yield term with a power law:

τ=τysgn⁡(γ˙)+K∣γ˙∣n−1γ˙.\tau=\tau_y\operatorname{sgn}(\dot{\gamma}) +K|\dot{\gamma}|^{n-1}\dot{\gamma}.

This allows yielding and shear thinning to coexist. Neither equation describes every detail of elastic deformation before flow, structural rebuilding, or the transition into motion.

Yielding produces distinctive velocity fields. In pressure-driven pipe flow, shear stress vanishes at the centerline and increases toward the wall. A yield-stress material can therefore contain an unyielded central plug moving with the surrounding sheared material. An unyielded region is not necessarily stationary: it may translate or rotate without continuously deforming internally. Conversely, some regions of a mixer can remain effectively immobilized if the local stresses never become sufficient to yield them.

Measured yield stresses depend on the protocol. A static yield stress, obtained after a specified rest period, can exceed the dynamic yield stress inferred from an already flowing material. Rest allows attractive particles, droplets, or other structural elements to rebuild load-bearing connections. The distinction is particularly important for materials whose structure changes with time.

An apparent threshold must also be distinguished from extremely slow flow. A finite-duration experiment may fail to detect creep occurring over hours or days. Wall slip can create the opposite difficulty: a sample may move against a smooth boundary even though its interior remains largely undeformed. Reliable interpretation consequently requires attention to observation time, instrument sensitivity, surface roughness, and local deformation. For some well-controlled materials, a yield-stress description is reproducible and useful; for others, the inferred threshold is inseparable from aging and measurement history.

Time dependence, thixotropy, and structural recovery

Rate dependence and time dependence are distinct. A purely shear-thinning liquid can, in an idealized description, attain a unique stress at each shear rate without any evolving internal structure. A thixotropic material instead undergoes a time-dependent reduction in flow resistance under sustained deformation, followed by structural recovery when deformation is reduced or stopped. Many clay dispersions, coatings, and structured food products display combinations of thixotropy, shear thinning, and yielding.

A simple conceptual model introduces a structural variable λ\lambda, often normalized between a relatively disrupted state and a more structured state. Its evolution may be written schematically as

dλdt=a(1−λ)−bλ∣γ˙∣m.\frac{d\lambda}{dt} =a(1-\lambda)-b\lambda|\dot{\gamma}|^m.

The first term represents rebuilding, while the second represents deformation-induced breakdown. Parameters aa, bb, and mm summarize material-specific kinetics rather than universal constants. Stress or viscosity is then made dependent on both λ\lambda and the instantaneous deformation rate. Such models express competition between structural recovery and disruption, although a single variable may be inadequate for materials containing several evolving structures.

A common experiment increases shear rate and then decreases it. If the sample has changed during the sweep, the descending stress curve may differ from the ascending curve, producing a hysteresis loop. However, the loop’s size and shape depend on sweep duration, rest history, and instrument response. Elastic transients, particle migration, and temperature changes can also generate hysteresis. A loop is therefore evidence of history dependence, not by itself a unique diagnosis of thixotropy.

The opposite time-dependent tendency—an increase in resistance during sustained shearing, conventionally called rheopexy or antithixotropy—is less common. Terminology varies, particularly when shear promotes an ordered or aggregated structure. Such behavior must be distinguished from instantaneous shear thickening: viscosity can increase with time at a fixed rate without following the same mechanism that causes a steady viscosity to increase with rate.

Some materials also undergo physical aging while resting. Their relaxation becomes progressively slower, and the stress required to restart flow can increase. An imposed deformation may partly erase this history, a process often termed rejuvenation. Recovery is not necessarily complete or reversible: irreversible aggregation, evaporation, chemical reactions, and damage can alter the sample permanently. These possibilities matter when interpreting repeated tests. A reproducible protocol must specify loading, preshear, rest duration, and measurement sequence, because ostensibly identical samples can otherwise begin an experiment in different structural states.

Viscoelasticity, normal stresses, and extensional flow

Viscoelasticity combines viscous flow with elastic storage of mechanical energy. A viscoelastic liquid can deform continuously under a sustained load yet recover part of its deformation when the load is removed. This behavior differs from a purely rate-dependent viscosity because stress retains information about previous deformation. Polymer solutions and melts are prominent examples, but micellar solutions and structured suspensions can also exhibit substantial elastic effects.

The linear Maxwell model places an ideal spring and an ideal viscous element in series. In simple shear its constitutive equation is

τ+λdτdt=ηγ˙,\tau+\lambda\frac{d\tau}{dt} =\eta\dot{\gamma},

where λ\lambda is a relaxation time. After a small, suddenly imposed strain that is then held fixed, the stress decays exponentially, with relaxation modulus

G(t)=G0e−t/λ,λ=ηG0.G(t)=G_0e^{-t/\lambda}, \qquad \lambda=\frac{\eta}{G_0}.

Real materials commonly possess a spectrum of relaxation times rather than one. The linear Maxwell equation describes small-deformation behavior; more elaborate, frame-independent equations are needed for general large deformations.

Timescale comparisons help organize these responses. The Deborah number compares a material relaxation time with an observation or process time,

De=λT,\mathrm{De}=\frac{\lambda}{T},

while the Weissenberg number commonly compares relaxation with a deformation rate, for example Wi=λγ˙\mathrm{Wi}=\lambda\dot{\gamma} in shear. Small values often correspond to ample time for relaxation. Large values indicate that substantial microstructural distortion or elastic memory may persist. The two numbers are related but answer different questions and need not coincide in a particular flow.

Viscoelastic fluids can also develop normal-stress differences. In simple shear, the first normal-stress difference is

N1=σxx−σyy.N_1=\sigma_{xx}-\sigma_{yy}.

Many polymer liquids develop positive N1N_1, helping produce rod climbing: liquid rises around a rotating shaft rather than displaying only the outward displacement familiar from inertia-dominated Newtonian flow. Elastic stress relaxation also contributes to die swell, in which an extruded stream expands after leaving a nozzle. These phenomena depend on flow geometry and the full stress state, not only on shear viscosity.

Stretching provides another distinction. For an incompressible Newtonian liquid in uniform uniaxial extension, the extensional viscosity is three times the shear viscosity. Polymer liquids can depart greatly from this ratio, exhibiting extension-rate dependence and strain hardening as chains become stretched. This affects filament breakup, spraying, fiber spinning, and droplet formation. Two formulations that appear similar in a rotational shear measurement may consequently produce very different threads, sprays, or extrusion defects.

Microscopic structures that produce complex flow

Non-Newtonian behavior arises when deformation changes structures that contribute significantly to stress. Relevant structures range from individual macromolecules to particle contacts and connected networks. Their characteristic sizes, interaction strengths, and relaxation times determine which responses appear under a given experimental condition. Chemical composition alone is therefore insufficient: the same ingredients can have different rheology when their concentration, particle size distribution, or preparation history differs.

In long-chain polymer liquids, molecular motion is constrained by neighboring chains. Above a concentration-dependent crossover, chains become entangled. Their stress relaxation involves motion within and eventual escape from these constraints, together with faster local motions. Increasing molecular mass can greatly increase viscosity and relaxation time in an entangled polymer melt. Branching, molecular-mass distribution, and chain stiffness further modify the response, especially during extension. A material can thus retain a molecular memory of deformation long after the deformation rate changes.

In colloidal dispersions, thermal agitation competes with interactions and imposed flow. Brownian motion tends to randomize particle arrangements, while attractive forces may produce clusters or system-spanning networks. Repulsive interactions can stabilize dispersions, but crowding at high concentration still restricts rearrangement. Flow distorts particle distributions, breaks aggregates, or creates contact networks. Larger, non-Brownian particles respond much less to thermal motion, making hydrodynamic forces, friction, and direct contacts especially important.

Interfacial structures provide another route. An emulsion consists of droplets dispersed in another liquid; when concentrated, droplets deform and crowd against one another. Foams behave similarly but contain gas bubbles separated by liquid films. Surface tension supplies a restoring force when droplets or bubbles are distorted. Their sizes, packing, interfacial chemistry, and susceptibility to coalescence influence elasticity and yielding. Two Newtonian constituent liquids can therefore form a strongly non-Newtonian emulsion.

Other systems obtain long-lived structure from reversible molecular association. Surfactant molecules may form elongated, wormlike micelles that entangle but can also break and recombine. Associative polymers form temporary junctions, while proteins and polysaccharides can build networks through multiple interactions. Hydrogen bonding, electrostatic forces, and hydrophobic association may all contribute, depending on the material.

Temperature, ionic conditions, and solvent quality change these structures as well as the viscosity of the surrounding liquid. For example, altering pH can change the charge and interactions of dispersed particles or macromolecules. Consequently, a rheological change need not indicate a single mechanical mechanism: it can reflect simultaneous changes in the continuous phase, interparticle interactions, and network connectivity.

Measurement methods and experimental limitations

Rheometry measures the relation between imposed deformation and material response. Rotational rheometers commonly use parallel plates, a cone and plate, or concentric cylinders. Depending on instrument design, they control rotational motion or applied torque and measure the complementary response. Geometry-dependent calculations convert these measurements into stress and deformation rate. Cone-and-plate arrangements approximate uniform shear rate under appropriate conditions, whereas parallel plates have a shear rate that varies with radius.

Steady-flow tests establish stress–rate curves, but time-resolved measurements reveal additional behavior. In a startup experiment, a constant shear rate is imposed and the stress is followed as the sample adjusts. A stress overshoot may reflect temporary loading followed by molecular orientation or structural disruption. In a cessation experiment, deformation stops and stress relaxation is monitored. Creep tests impose a stress and measure deformation; subsequent recovery helps distinguish recoverable deformation from accumulated flow.

Small-amplitude oscillatory shear probes linear viscoelasticity by applying

γ(t)=γ0sin⁡(ωt).\gamma(t)=\gamma_0\sin(\omega t).

The stress response can be expressed as

τ(t)=γ0[G′(ω)sin⁡(ωt)+G′′(ω)cos⁡(ωt)].\tau(t)=\gamma_0 \left[G'(\omega)\sin(\omega t) +G''(\omega)\cos(\omega t)\right].

The storage modulus G′G' characterizes the in-phase elastic contribution, while the loss modulus G′′G'' characterizes the out-of-phase dissipative contribution. Measurements at several amplitudes establish the linear-response range; frequency sweeps then probe different timescales. A crossover between the moduli can indicate a characteristic relaxation scale, but assigning one molecular mechanism to it requires additional evidence.

Capillary rheometers are useful at high flow rates and for polymer processing. Interpreting their pressure–flow data may require corrections for entrance losses, nonuniform shear rate, and wall slip. Extensional rheometers and filament-thinning methods examine stretching, although maintaining homogeneous extension can be difficult. Imaging and local velocimetry—including ultrasound, optical methods, and magnetic resonance imaging—can reveal flow profiles hidden by bulk measurements.

Major artifacts include wall slip, evaporation, sedimentation, particle migration, viscous heating, and sample fracture or edge instability. Roughened boundaries can reduce slip, but an inappropriate gap or roughness scale introduces other errors. At high rates, inertia can influence the measured response; at low stresses, torque resolution and mechanical drift become critical. Temperature control is especially important because dissipation can warm a sample and change its viscosity during measurement.

Calibration with suitable standards and a stated measurement uncertainty support reliable comparisons. For history-dependent materials, however, instrument accuracy is only one part of reproducibility. Sample composition, age, handling, preshear, rest time, and test duration are also essential experimental information.

Historical development, mathematical models, and computation

The Newtonian reference law derives from Isaac Newton’s treatment of fluid resistance in the Principia of 1687. Its usefulness does not imply that all liquids obey it under all conditions. As industrial chemistry and materials processing expanded, measurements of suspensions, rubbers, pastes, and polymeric liquids required descriptions beyond constant viscosity. The field developed through interaction between practical processing problems and increasingly detailed theories of matter.

Nineteenth-century elasticity and viscosity models supplied foundations for describing memory. James Clerk Maxwell developed an influential relaxation model in the 1860s, linking transient stress to elastic and viscous behavior. In the early twentieth century, Eugene C. Bingham helped establish quantitative descriptions of plastic flow and yield-stress materials. The Society of Rheology was founded in 1929, institutionalizing a discipline that encompassed liquids, soft solids, and materials lying between conventional categories.

Mid-twentieth-century constitutive theories sought equations valid for large, three-dimensional deformations. James G. Oldroyd’s 1950 work on rheological equations of state was influential in formulating objective stress evolution: predictions should not change merely because an observer changes reference frame. Models such as Oldroyd-B combine a Newtonian solvent contribution with an idealized polymeric stress. They capture important elastic effects but can fail quantitatively in strong flows, including situations involving finite molecular extensibility.

At a more microscopic level, theories associated with Pierre-Gilles de Gennes, Masao Doi, and Sam Edwards connected polymer relaxation to chain constraints and entanglements. Other developments related suspension stresses to hydrodynamic interactions, particle forces, and evolving microstructure. These approaches showed why one empirical flow curve cannot identify a unique internal mechanism. They also established routes for predicting how molecular architecture or particle interactions affect measurable properties.

In computation, a constitutive model is coupled to the equations expressing conservation of mass and momentum balance. Generalized Newtonian models use a viscosity depending on a deformation-rate invariant and are often practical when memory and normal stresses are secondary. Viscoelastic simulations evolve additional stress or conformation fields, while structural models introduce variables for aggregation, aging, or contact networks. Yield-stress calculations may require explicit treatment of unyielded regions or regularization of an ideal threshold.

Important numerical difficulties include steep stress gradients, moving yielded boundaries, and instability at high Weissenberg number. Microscale methods, including molecular dynamics and particle-based simulations, can investigate mechanisms but generally cannot reproduce every industrial length and timescale directly. Model selection therefore involves a controlled compromise: an equation must represent the physical effects relevant to the application, remain mathematically and numerically tractable, and be tested against measurements beyond those originally used to fit its parameters.

Biological, geological, and industrial significance

Blood is a prominent biological non-Newtonian fluid. Its behavior depends strongly on the concentration, aggregation, and deformability of red blood cells. At low shear rates, reversible cellular aggregation can increase apparent viscosity; at higher rates, aggregate disruption and cellular deformation contribute to shear thinning. In sufficiently small vessels, cell size becomes comparable to the flow geometry, and a homogeneous continuum description becomes less adequate. Blood viscosity is therefore not a single number independent of vessel size and flow conditions.

Mucus provides a different example: hydrated macromolecular networks confer viscoelasticity and interact with particles and surfaces. Biological gels and the extracellular matrix similarly exhibit time-dependent mechanics arising from polymer networks, reversible connections, and fluid redistribution. These responses are not all reducible to the same fluid model. Some biological materials behave predominantly as elastic solids over relevant observation times, while others flow or remodel substantially.

In the environment, concentrated sediment suspensions, mudflows, and some debris flows exhibit yielding and rate-dependent resistance. Their mobility depends on water content, particle concentration, grain-size distribution, and the balance between fine-particle cohesion and coarse-grain contacts. A debris flow can contain fluid, sediment, and large clasts whose interactions require multiphase descriptions. Representing every such flow as a homogeneous Bingham fluid can obscure segregation, pore-pressure evolution, and granular friction.

Non-Newtonian behavior is equally important in manufacturing. Paint formulations often balance resistance to sagging at rest with easy spreading during application. Drilling fluids must transport cuttings and limit their settling during interruptions, linking low-rate structure to high-rate pumpability. Fresh concrete requires a balance between yield stress, plastic viscosity, and resistance to segregation. In food processing, rheology influences mixing, filling, texture, and the stability of suspended ingredients, while the effects of temperature and composition must be separated from those of deformation history.

Polymer extrusion, film manufacture, and fiber spinning depend strongly on elastic and extensional stresses. These stresses can support a stretched filament but can also contribute to instabilities and surface defects. Small amounts of dissolved high-molecular-mass polymer can reduce turbulent drag in suitable flows, showing that elastic microstructure can alter large-scale motion even when the solution’s shear viscosity changes only modestly.

Material design increasingly exploits such coupled responses. Printable inks may be formulated to flow through a nozzle and regain enough structure afterward to retain shape. Dense shear-thickening suspensions have been investigated for impact-responsive composites, although performance depends on confinement, loading rate, and the surrounding structure. Across these applications, useful specifications involve a range of stresses, deformation rates, and recovery times rather than a single label such as “thick,” “elastic,” or “non-Newtonian.”