Viscosity
Viscosity is a measure of a fluid's resistance to flow, describing the internal friction that opposes the relative motion of its layers.[1] This property arises from interactions between fluid molecules, transforming kinetic energy of bulk motion into thermal energy through dissipative processes akin to friction.[2] In fluids, viscosity governs behaviors such as the spreading of liquids, the flow through pipes, and the formation of boundary layers in aerodynamics, making it a fundamental parameter in fluid dynamics.[3] Viscosity is quantified in two primary forms: dynamic viscosity (also called absolute viscosity) and kinematic viscosity.[3] Dynamic viscosity, denoted by the symbol μ, represents the fluid's resistance to shear stress under an applied force and is defined by Newton's law of viscosity as the ratio of shear stress (τ) to the velocity gradient (du/dy) in laminar flow: τ = μ (du/dy).[3] Kinematic viscosity, denoted by ν, is the dynamic viscosity divided by the fluid's density (ρ), or ν = μ / ρ, and is commonly used to describe flows influenced by body forces such as gravity.[3] The SI unit for dynamic viscosity is the pascal-second (Pa·s), equivalent to one newton-second per square meter (N·s/m²), while kinematic viscosity is measured in square meters per second (m²/s).[2] Viscosity varies significantly with temperature—for liquids, it decreases as temperature rises due to reduced molecular cohesion, whereas in gases, it increases with temperature because of higher molecular speeds.[2] These dependencies are critical in applications ranging from lubrication in engines to atmospheric modeling. In engineering and physics, viscosity influences phenomena like drag forces, turbulence transitions via the Reynolds number (Re = ρVL / μ), and material processing, where high-viscosity fluids like honey exhibit slow deformation compared to low-viscosity ones like water.[3] Understanding viscosity enables predictions of fluid behavior in diverse fields, including aerospace, chemical engineering, and geophysics.[3]Etymology and Definitions
Etymology
The word "viscosity" derives from the Late Latin viscositas, meaning "stickiness," which stems from viscosus ("sticky") and ultimately from the Latin viscum ("mistletoe"). This etymological root refers to the viscous, adhesive birdlime produced from the berries of the mistletoe plant, used historically to trap birds, and the term was later extended metaphorically to characterize the clinging or resistant quality of semi-fluid substances.[4][5] Early conceptual notions of viscosity trace back to ancient natural philosophy, where qualitative descriptions of fluid resistance appeared without the modern term. For instance, Aristotle, in his Physics (circa 350 BCE), observed that objects moving through denser media like water experience greater opposition to motion than in air, attributing this to the medium's inherent resistance that scales with its "thickness."[6] The English term "viscosity" itself emerged in the late 14th century, initially denoting a general state of viscidity or glutinousness, and entered scientific usage in the 17th century amid investigations into fluid behavior. Isaac Newton's Philosophiæ Naturalis Principia Mathematica (1687) formalized proportional relationships in fluid resistance, influencing later terminology without explicitly defining viscosity as a distinct property.[4] In the 19th century, the term gained precise application in continuum mechanics when Claude-Louis Navier incorporated viscous effects into his 1822 generalization of Leonhard Euler's inviscid fluid equations, marking a pivotal formalization of viscosity as internal fluid friction, though the word had long predated this in broader natural philosophical contexts.[7]Dynamic Viscosity
Dynamic viscosity, denoted by the symbol μ, is a fundamental property of fluids that quantifies their resistance to flow under applied shear forces. It is defined as the ratio of the shear stress τ to the shear rate, which is the velocity gradient du/dy perpendicular to the flow direction.[8][9] This relationship is expressed mathematically as \mu = \frac{\tau}{\frac{du}{dy}} where τ represents the shear stress, the internal frictional force per unit area that opposes the relative motion of adjacent fluid layers sliding past one another.[8][9] In simple shear flow, such as between two parallel plates where one moves relative to the other, this gradient du/dy arises from the velocity difference across the fluid layers, and dynamic viscosity characterizes the linear proportionality between this stress and rate for Newtonian fluids.[10] Unlike bulk viscosity, which pertains to a fluid's resistance to uniform compression or expansion, dynamic viscosity specifically addresses deformation due to shear without volume change, making it central to understanding frictional effects in laminar flows.[11] This distinction ensures that dynamic viscosity focuses on tangential stresses in directional flows, excluding dilatational contributions.[11] The concept originates from Isaac Newton's 1687 work Philosophiæ Naturalis Principia Mathematica, where he postulated that for ideal fluids, the resistive force is proportional to velocity and independent of the shear rate, assuming constancy of viscosity at a given temperature—a foundational assumption for Newtonian behavior.[12][13] This linearity underpins the definition, distinguishing it from more complex fluid responses.[13]Kinematic Viscosity
Kinematic viscosity, denoted by the symbol \nu, is defined as the ratio of a fluid's dynamic viscosity \mu to its density \rho, expressed as \nu = \frac{\mu}{\rho}.[3] This measure has units of area per unit time, such as square meters per second in the SI system.[14] Physically, kinematic viscosity represents the diffusivity of momentum within the fluid, quantifying how readily momentum is transported through the medium due to viscous effects.[15] It characterizes the fluid's resistance to shear relative to its inertial properties, making it particularly useful in analyses where density variations influence flow behavior.[3] A key application of kinematic viscosity is in the Reynolds number, Re = \frac{v L}{\nu}, where v is the flow velocity and L is a characteristic length; this dimensionless quantity predicts whether a flow will be laminar or turbulent by comparing inertial to viscous forces.[3][14] The value of kinematic viscosity varies with temperature and pressure through the dependencies of both \mu and \rho; for instance, in ideal gases, \nu scales proportionally to T^{3/2}/p, where T is temperature and p is pressure, since dynamic viscosity \mu increases with temperature while being largely independent of pressure, and density \rho decreases with rising temperature or falling pressure.[16][15] In liquids, temperature typically reduces \mu more significantly, leading to a decrease in \nu, whereas pressure effects on \rho are more pronounced but often secondary.[3]Fluid Behavior
Newtonian Fluids
A Newtonian fluid is defined by the linear relationship between shear stress \tau and the rate of strain \dot{\gamma}, expressed as \tau = \eta \dot{\gamma}, where \eta is the dynamic viscosity that remains constant and independent of the shear rate \dot{\gamma}.[17] This constitutive relation, known as Newton's law of viscosity, implies that the fluid's resistance to flow does not vary with the intensity of deformation, distinguishing it from more complex behaviors in other fluids.[18] Common examples of Newtonian fluids include water, air, glycerine, and most simple gases and low-molecular-weight liquids under typical conditions of low shear.[19] These fluids exhibit constant viscosity across a wide range of shear rates, making their flow behavior predictable in engineering applications such as pipelines and lubrication systems.[20] The constant viscosity of Newtonian fluids has key implications for flow dynamics, particularly in enabling laminar flow at low Reynolds numbers, where fluid layers maintain smooth, parallel motion without mixing or turbulence.[21] For instance, the drag force F_d on a small sphere of radius r moving at low velocity v through such a fluid is given by Stokes' law: F_d = 6 \pi \eta r v, which quantifies viscous resistance in sedimentation and colloidal systems.[22] This linearity arises from fundamental physical mechanisms. In dilute gases, kinetic theory explains viscosity as the net momentum transfer between adjacent layers by molecules with average thermal speed \bar{v} traversing the mean free path \lambda, yielding \eta \approx \frac{1}{3} \rho \lambda \bar{v}, where \rho is density; the proportionality to the velocity gradient ensures \eta remains independent of \dot{\gamma}.[23] For low-molecular-weight liquids, short-range intermolecular forces dominate, providing a consistent frictional resistance without rate-dependent alignment or entanglement effects.[24]Non-Newtonian Fluids
Non-Newtonian fluids are substances in which the viscosity does not remain constant but varies with the applied shear rate or over time, deviating from the linear relationship between shear stress and shear rate observed in Newtonian fluids.[25] This behavior arises in complex fluids containing particles, polymers, or other microstructures that respond to deformation.[26] Non-Newtonian fluids are broadly classified into time-independent and time-dependent categories based on how their flow resistance changes. In time-independent non-Newtonian fluids, viscosity depends solely on the instantaneous shear rate. Shear-thinning fluids, also known as pseudoplastic, exhibit decreasing viscosity as shear rate increases, allowing easier flow under stress; common examples include paints and polymer solutions. Conversely, shear-thickening or dilatant fluids show increasing viscosity with higher shear rates, often due to structural alignment or particle interactions; a classic example is a cornstarch-water slurry, which hardens under rapid impact.[25] Bingham plastics represent yield-stress fluids that behave as solids below a critical shear stress but flow like viscous liquids above it, exemplified by toothpaste, which holds its shape until squeezed.[26] Time-dependent non-Newtonian fluids, such as thixotropic ones, display viscosity that decreases over time under constant shear due to reversible structural breakdown, with recovery upon rest; this is seen in certain inks and gels.[25] A key model for describing shear-thinning and shear-thickening behaviors is the Ostwald-de Waele power-law model, which relates shear stress \tau to shear rate \dot{\gamma} as: \tau = K \dot{\gamma}^n where K is the consistency index and n is the flow behavior index; n < 1 indicates shear-thinning, while n > 1 denotes shear-thickening.[27] The apparent viscosity \eta_\text{app} is then defined as \eta_\text{app} = \tau / \dot{\gamma} = K \dot{\gamma}^{n-1}, which varies with shear rate, providing a measure of effective flow resistance.[28] Everyday relevance of non-Newtonian fluids is evident in biological and industrial contexts. Blood, for instance, acts as a shear-thinning fluid, with viscosity dropping at higher shear rates in vessels to facilitate circulation.[29] Polymer melts and solutions in manufacturing also exhibit pseudoplastic behavior, enabling efficient processing in extrusion and coating applications. These properties influence product design, from non-drip paints to protective body armors using dilatant materials.[25]Physical Mechanisms
Momentum Transport
Viscosity manifests as the transport of momentum across adjacent fluid layers that exhibit relative motion, arising from intermolecular forces that enable momentum exchange perpendicular to the primary flow direction. In a shearing flow, molecules from a faster-moving layer collide with those in a slower layer, effectively diffusing momentum downward and equalizing velocities over time.[30] This process is analogous to diffusion phenomena, where viscosity acts as the diffusivity for momentum, smoothing out velocity gradients and resisting shear.[31] A classic illustration occurs in laminar flow through a channel, such as Poiseuille flow between parallel plates, where the momentum transport due to viscosity results in a parabolic velocity profile. The maximum velocity is at the center, decreasing symmetrically toward the walls due to the no-slip condition, with the profile governed by the balance between pressure-driven advection and viscous diffusion of momentum.[30] This parabolic shape emerges directly from solving the simplified Navier-Stokes equations under steady, incompressible conditions, highlighting how viscous momentum transfer enforces the velocity variation.[32] The quantitative relation for this transport in simple shear is captured by Newton's law of viscosity, expressed as the shear stress \tau = -\eta \frac{du}{dy}, where \eta is the dynamic viscosity and \frac{du}{dy} is the velocity gradient normal to the flow.[30] This form parallels Fick's first law for mass diffusion (J = -D \frac{dc}{dy}) and Fourier's law for heat conduction (q = -k \frac{dT}{dy}), underscoring the unified framework of transport processes in continua, with \eta serving as the momentum diffusivity coefficient.[31] Within the full Navier-Stokes equations, viscous effects enter through the divergence of the stress tensor, specifically the term \nabla \cdot (\eta \nabla \mathbf{v}) (for constant \eta), which represents the net diffusive flux of momentum and acts to dampen velocity fluctuations.[32] This term is essential for describing the evolution of velocity fields in viscous flows. In practical applications, such momentum transport dominates in boundary layers adjacent to solid surfaces, where it generates skin friction drag by slowing fluid near the wall and creating low-momentum regions.[33] Strategies for drag reduction, such as injecting low-viscosity fluids or using polymer additives, target these layers to enhance momentum transfer away from the wall, thereby reducing shear stress and overall resistance.[34]Molecular Origins in Gases
In dilute gases, viscosity originates from the diffusive transport of momentum between adjacent fluid layers through intermolecular collisions, as described by kinetic theory. Molecules traveling across velocity gradients carry excess momentum from faster-moving layers to slower ones, resulting in a net shear stress that opposes the flow. This microscopic mechanism underpins the macroscopic viscous resistance observed in gases.[35] The foundational estimate from kinetic theory for the shear viscosity \eta of a dilute gas is given by \eta \approx \frac{1}{3} \rho \lambda v_{\rm avg}, where \rho is the mass density, \lambda is the mean free path between collisions, and v_{\rm avg} is the average molecular speed. This expression arises from considering the flux of molecules across a plane, each transporting momentum on the order of m v_z v_x (with m the molecular mass and v_x, v_z velocity components), averaged over the Maxwell-Boltzmann distribution and assuming a velocity gradient over the scale of \lambda. James Clerk Maxwell derived this form in his pioneering work on gas friction, confirming its proportionality to molecular speed and path length.[35][36] A key feature of this model is the independence of \eta from gas density (or pressure) at constant temperature, valid for dilute conditions where the mean free path is much larger than molecular sizes. The mean free path \lambda \approx 1/(\sqrt{2} \pi d^2 n) inversely scales with number density n (and thus \rho = m n), so the product \rho \lambda remains constant, yielding \eta independent of n. This counterintuitive result, predicted by Maxwell, was experimentally verified and distinguishes gaseous viscosity from that in denser fluids. With temperature, v_{\rm avg} \propto \sqrt{T} from equipartition, while \lambda is temperature-independent in the basic model, leading to \eta \propto T^{1/2}.[35][37] The simple hard-sphere model, assuming molecules as rigid spheres of diameter d with no long-range forces, provides a good first approximation for monatomic gases like helium or argon at moderate conditions but has limitations. It underpredicts the temperature dependence (T^{1/2}) compared to experiments, where real gases show a stronger increase due to attractive intermolecular potentials that effectively reduce the collision cross-section at higher temperatures. Corrections for real gases, such as the Sutherland model, modify the effective collision cross-section \sigma \propto d^2 (1 + S/T), where S is a characteristic temperature reflecting attractive forces, yielding \eta \propto T^{1/2} / (1 + S/T). This semi-empirical adjustment, originally proposed by William Sutherland, improves accuracy for polyatomic gases like air over wide temperature ranges without resorting to full quantum treatments.[36][38] For bulk viscosity \zeta, which arises in compressible flows involving volume dilation, kinetic theory predicts \zeta \approx 0 for dilute monatomic gases. Without internal degrees of freedom (e.g., rotation or vibration), there is no relaxation time for energy redistribution during compression or expansion, so no additional dissipative resistance beyond shear effects. This result holds in the hard-sphere approximation and is confirmed by the Chapman-Enskog solution for low-density monatomic gases.[39]Molecular Origins in Liquids
In liquids, viscosity arises primarily from the strong intermolecular interactions and the cooperative motion required for molecules to flow past one another in a dense medium, contrasting with the dilute collision-dominated transport in gases. Unlike gases, where viscosity stems from momentum transfer via infrequent binary collisions, liquids exhibit viscosity due to the caged dynamics of molecules surrounded by neighbors, leading to higher resistance to shear as density increases. This dense-phase behavior results in activation barriers for flow, where molecules must overcome potential energy hurdles to rearrange, often modeled through extensions of kinetic theory adapted for correlated motions.[40] The Enskog theory, originally developed for dense gases, has been extended to liquids by incorporating corrections for frequent collisions and spatial correlations in high-density regimes, predicting that viscosity η increases with density due to enhanced collision rates and reduced mean free paths. In these modifications, such as the modified Enskog theory (MET), the viscosity is expressed as η = η_0 * Y, where η_0 is the low-density limit from Chapman-Enskog theory, and Y is a density-dependent Enskog correction factor accounting for pair correlations that amplify momentum transfer in crowded environments. This extension successfully describes how higher liquid densities elevate viscosity by promoting more entangled molecular trajectories, as validated for simple fluids like argon near saturation.[41][42] Temperature dependence in liquids follows an Arrhenius-like form, η = A exp(E_a / RT), where A is a pre-exponential factor, E_a is the activation energy for viscous flow reflecting the energy barrier to molecular rearrangement, R is the gas constant, and T is absolute temperature; this exponential increase in viscosity with decreasing temperature underscores the role of thermal energy in overcoming intermolecular attractions. For many organic liquids, E_a correlates with molecular size and polarity, typically ranging from 10-30 kJ/mol, as derived from empirical fits to experimental data. This model holds well above the glass transition but deviates at lower temperatures where cooperative effects dominate.[43][44] Intermolecular forces significantly influence liquid viscosity, with hydrogen bonding in water creating a dynamic network that enhances resistance to flow by forming transient bridges between molecules, leading to water's anomalously high viscosity compared to non-hydrogen-bonding liquids of similar mass. In contrast, van der Waals forces dominate in nonpolar oils, where dispersion interactions between hydrocarbon chains increase viscosity proportional to chain length and branching, as longer chains foster greater entanglement and slower relaxation times. These forces contribute to the scale of viscosity, with hydrogen-bonded systems like water showing E_a ≈ 16 kJ/mol, while van der Waals-dominated oils like decane exhibit higher values around 25 kJ/mol.[45][46] Free volume theory provides a complementary perspective, positing that viscosity rises dramatically as available free volume—the unoccupied space per molecule—decreases near the glass transition temperature T_g, where molecular mobility freezes due to insufficient space for diffusive jumps. In this framework, the diffusion coefficient D ∝ exp(-B / v_f), with v_f the fractional free volume and B a constant, linking viscosity inversely to D via the Stokes-Einstein relation; as temperature drops below T_g + 50 K, v_f shrinks, causing η to span orders of magnitude from 10^2 Pa·s at T_g to 10^{12} Pa·s in the glassy state. This theory, pioneered by Cohen and Turnbull, explains the universal super-Arrhenius behavior in dense liquids approaching vitrification.[47][48]Extensions to Other Systems
Bulk Viscosity
Bulk viscosity, denoted as \zeta, represents the fluid's resistance to uniform volumetric compression or expansion, in contrast to dynamic viscosity which governs resistance to shear deformation. In the context of compressible fluid flow, it manifests as a deviation in the mechanical pressure from its thermodynamic equilibrium value, related to the divergence of the velocity field through the constitutive relation P' = -\zeta (\nabla \cdot \mathbf{v}), where P' is the pressure deviation and \mathbf{v} is the velocity vector. This term arises in the Navier-Stokes equations for the viscous stress tensor, capturing dissipative effects during isotropic volume changes. In simple fluids such as monatomic gases in the dilute limit, bulk viscosity is negligible, approximately \zeta \approx 0, because these systems lack internal degrees of freedom or significant intermolecular forces that could lead to delayed equilibration during compression or expansion. However, in polyatomic gases and molecular liquids, bulk viscosity becomes substantial due to relaxation processes involving internal modes, such as rotational and vibrational excitations, which cannot instantaneously adjust to rapid volume changes, resulting in non-equilibrium pressure contributions. For instance, in polyatomic species like nitrogen, these mechanisms introduce a finite \zeta that scales with the complexity of molecular structure.[49] Bulk viscosity plays a key role in acoustic propagation and shock wave dynamics. It contributes to sound absorption, where the classical attenuation coefficient includes a term proportional to \zeta, specifically \alpha = \frac{\omega^2}{2\rho c^3} \left( \frac{4}{3} \eta + \zeta + (\gamma - 1)^2 \frac{\kappa}{\rho c_p} \right), with \eta as dynamic viscosity, \kappa thermal conductivity, \rho density, c sound speed, \gamma the adiabatic index, and \omega angular frequency; the \zeta term accounts for structural relaxation damping. In shock waves, bulk viscosity influences the transition zone thickness, with higher \zeta leading to broader profiles; theoretical models show a linear dependence of the normalized shock thickness on the ratio \zeta / \eta, as observed in experiments with polyatomic gases like sulfur hexafluoride. Experimental determination of bulk viscosity primarily relies on ultrasound attenuation measurements, which isolate the excess absorption not explained by shear viscosity or thermal conduction. Acoustic spectroscopy techniques have been applied to various Newtonian fluids, including gases and liquids, revealing \zeta values independent of properties like density or shear viscosity; for example, in liquid water and organic solvents, attenuation data yield \zeta on the order of 10^{-3} to 10^{-2} Pa·s across frequencies from 1 to 100 MHz.[50] Similar ultrasound experiments in polyatomic gases, such as nitrogen from 77 K to 300 K, confirm the role of vibrational relaxation in generating measurable \zeta, with absorption spectra fitting models that attribute up to 50% of total damping to bulk effects.[49]Viscosity in Solids
In solids, viscosity manifests as resistance to deformation rates under sustained stress, analogous to fluid shear viscosity but applied to the creep behavior of viscoelastic materials. In creep tests, where a constant stress σ is applied, the resulting strain ε(t) increases over time, and the creep compliance is defined as J(t) = ε(t) / σ, quantifying the material's time-dependent deformability.[51] The Maxwell model, a fundamental viscoelastic framework consisting of a spring (elastic modulus E) in series with a dashpot (viscosity η), captures this by relating the viscous component to the strain rate: the dashpot alone yields dε/dt = σ / η, so η = σ / (dε/dt), where the total strain rate combines elastic and viscous contributions. This model highlights how solids can exhibit fluid-like flow under prolonged loading, distinguishing viscosity from pure elasticity, which involves instantaneous, recoverable deformation without time dependence.[52] Unlike ideal elastic solids, many real solids possess a yield stress beyond which permanent deformation occurs, yet they demonstrate viscous flow characteristics particularly at elevated temperatures or over extended timescales. For instance, glacier ice deforms viscously under its own weight, flowing like a non-Newtonian fluid with an effective viscosity on the order of 10^{13} to 10^{14} Pa·s, enabling slow movement over geological periods without fracturing.[53] Similarly, pitch, a highly viscous amorphous solid at room temperature with η ≈ 10^{8} Pa·s, flows imperceptibly slowly, as demonstrated by the ongoing pitch drop experiment where drops form roughly every decade, illustrating solid-like rigidity masking underlying viscous behavior.[54] Amorphous solids, such as glassy polymers, can be viewed as supercooled liquids frozen into a non-equilibrium state, where viscous flow dominates near the glass transition temperature T_g.[55] In glassy polymers, viscosity reaches extraordinarily high values near T_g, typically around 10^{12} Pa·s, marking the boundary where structural relaxation times approach observable scales (e.g., 100 seconds), transitioning the material from a viscous liquid to a rigid glass.[56] This regime is critical for applications like polymer processing, where flow resistance governs shaping and annealing. Recent advances in 2025 have introduced atomistic models that connect near-glass-transition viscosity directly to the full spectrum of atomic vibration modes, using non-affine lattice dynamics to compute shear viscosity from low-frequency vibrational contributions without relying on computationally intensive simulations. These models, validated on polymer melts like the Kremer-Grest system, reveal how collective vibrational anharmonicity enhances flow resistance, providing predictive power across temperatures where traditional methods falter.[57]Eddy Viscosity
Eddy viscosity, denoted as \nu_t, is an empirical concept in turbulence modeling that parameterizes the enhanced momentum transport due to turbulent eddies, analogous to molecular viscosity in laminar flows. It is defined by the relation \nu_t = \tau_t / [\rho (\partial u / \partial y)], where \tau_t represents the turbulent shear stress, \rho is the fluid density, and \partial u / \partial y is the mean velocity gradient in the direction perpendicular to the flow. This formulation arises from the Boussinesq hypothesis, which assumes that turbulent fluctuations act like an additional viscous stress on the mean flow.[58] A foundational approach to estimating eddy viscosity is Prandtl's mixing-length theory, introduced in the early 20th century. According to this theory, \nu_t \approx l^2 |\partial u / \partial y|, where l is the mixing length, a characteristic scale representing the average distance traveled by turbulent eddies before their momentum is redistributed. Near walls, l is often taken as proportional to the distance from the surface, such as l = \kappa y with von Kármán constant \kappa \approx 0.41. This simple model effectively captures the shear stress in flows dominated by a single length scale.[59][58] In practice, eddy viscosity plays a central role in closing the Reynolds-Averaged Navier-Stokes (RANS) equations for simulating turbulent flows in engineering applications, such as fully developed pipe flows and atmospheric boundary layers. By incorporating \nu_t into the effective viscosity, RANS models approximate the Reynolds stresses as \tau_{ij} = \rho \nu_t (\partial U_i / \partial x_j + \partial U_j / \partial x_i) - (2/3) \rho k \delta_{ij}, enabling computationally efficient predictions of mean flow fields without resolving individual eddies. These models are particularly valuable for design in aerospace, civil engineering, and meteorology, where high-fidelity direct simulations are infeasible.[60][59] However, eddy viscosity is not a true thermophysical property of the fluid like molecular viscosity; instead, it varies spatially and temporally with local flow conditions, turbulence intensity, and geometry, often requiring ad hoc tuning or additional transport equations for its prediction. This dependence leads to limitations in non-equilibrium flows, such as those with strong streamline curvature, separation, or rapid distortions, where the isotropic assumption fails and models can produce unphysical results like excessive diffusion or instabilities.[58][61]Measurement and Units
Measurement Techniques
Viscosity measurement techniques vary depending on the fluid type, viscosity range, and whether the fluid behaves as Newtonian or non-Newtonian, with methods designed to apply controlled shear and quantify resistance to flow. Common approaches include capillary, rotational, falling sphere, and oscillatory methods, each leveraging fundamental fluid dynamics principles to derive viscosity from measurable parameters like pressure drop, torque, or velocity. These techniques are calibrated against standard fluids to ensure accuracy, often achieving precisions better than 1% for low-viscosity liquids.[62] Capillary viscometers are widely used for low-viscosity Newtonian fluids, such as gases and light oils, by forcing the fluid through a narrow tube and measuring the flow rate under a pressure difference. The method relies on the Hagen-Poiseuille equation, which assumes laminar, fully developed flow in a cylindrical capillary: \Delta P = \frac{8 \eta L Q}{\pi r^4}, where \Delta P is the pressure drop, \eta is the dynamic viscosity, L is the capillary length, Q is the volumetric flow rate, and r is the radius; solving for \eta yields viscosity directly from experimental measurements of \Delta P and Q. This technique is effective for viscosities ranging from 0.1 to 100 mPa·s, with automated versions enabling high-throughput analysis in pharmaceutical applications.[62][63] Rotational viscometers, particularly those employing Couette geometry with coaxial cylinders, measure viscosity by rotating an inner cylinder within a stationary outer one filled with the fluid and quantifying the resulting torque. For Newtonian fluids in this setup, the torque T balances the viscous shear, given by T = \frac{4 \pi \eta \Omega h r_i^2 r_o^2}{r_o^2 - r_i^2}, where \Omega is the angular velocity, h is the cylinder height (or immersion length), and r_i, r_o are the inner and outer radii, respectively; viscosity is then computed from measured torque and rotation speed. This method suits moderate to high viscosities (up to 10^6 mPa·s) and can handle opaque samples, making it versatile for industrial fluids like paints and polymers./20%3A_Miscellaneous/20.04%3A_Viscosity/20.4.02%3A_The_Couette_Viscometer) The falling sphere viscometer determines viscosity by observing the terminal velocity of a sphere descending through a transparent fluid column under gravity, applicable to low-viscosity Newtonian liquids like water or oils. Based on Stokes' law for the drag force F_d = 6 \pi \eta r v balancing the buoyant weight, the terminal velocity v satisfies \eta = \frac{2 r^2 g (\rho_s - \rho_f)}{9 v}, where r is the sphere radius, g is gravitational acceleration, and \rho_s, \rho_f are the densities of the sphere and fluid; high-speed imaging or timing tracks v for precise calculation. This technique is simple and absolute, offering accuracies of 0.5-2% for viscosities below 100 mPa·s, though wall effects require corrections for finite tube diameters.[64][65] Oscillatory rheometers extend viscosity assessment to non-Newtonian and high-viscosity materials, such as gels and viscoelastic polymers, by applying sinusoidal shear strain and measuring the stress response to derive dynamic moduli. The loss modulus G'' relates to the viscous component via the dynamic viscosity \eta' = G'' / \omega, where \omega is the angular frequency, allowing characterization of shear-rate-dependent behavior without excessive deformation. These instruments, often using parallel-plate or cone-plate geometries, operate in the linear viscoelastic regime (strain < 1%) and are essential for complex fluids where steady shear may induce structural changes.[66] Recent advances in microfluidic devices have enabled viscosity measurements on microliter-scale samples, particularly for biological fluids and high-value materials, overcoming limitations of traditional methods in sample volume and portability. For instance, chip-based viscometers integrate pressure-driven flow in microchannels to apply Poiseuille-like principles, achieving rapid (seconds) assessments with 10 μL volumes and accuracies of 5% for protein solutions up to 10 mPa·s. These post-2020 innovations, including temperature-controlled rheometers, facilitate in situ monitoring of non-Newtonian effects like shear thinning in complex mixtures.[67][68]Units and Dimensions
The standard unit for dynamic viscosity in the International System of Units (SI) is the pascal-second (Pa·s), which is equivalent to the newton-second per square meter (N·s/m²).[69] In the centimeter-gram-second (CGS) system, the corresponding unit is the poise (P), defined as the dyne-second per square centimeter (dyne·s/cm²).[70] The conversion between these units is given by 1 P = 0.1 Pa·s.[70] Kinematic viscosity, defined as the ratio of dynamic viscosity to fluid density, has the SI unit of square meters per second (m²/s).[70] In the CGS system, it is measured in stokes (St), where 1 St = 10^{-4} m²/s.[70] Dimensional analysis yields the expression for the dimension of dynamic viscosity as [\eta] = M L^{-1} T^{-1}, where M represents mass, L length, and T time.[71] Historical units for viscosity, particularly in the petroleum industry, include Saybolt Universal Seconds (SUS), which quantify the time in seconds for a fixed volume of fluid to flow through a standardized orifice and are primarily applied to oils.[72]Theoretical Prediction
Chapman–Enskog Theory for Gases
The Chapman–Enskog expansion provides a systematic perturbation method to solve the Boltzmann equation for dilute gases, deriving expressions for transport coefficients such as viscosity directly from the intermolecular potential function. Developed independently by Sydney Chapman in 1916–1917 and David Enskog in 1917, and later refined in their collaborative work, the theory expands the velocity distribution function in powers of the Knudsen number (the ratio of mean free path to macroscopic length scale), assuming small gradients in velocity, temperature, and density. To first order, this yields the Navier–Stokes constitutive relations, with viscosity expressed as a function of molecular mass, temperature, and collision parameters derived from the potential. The first-order approximation for the shear viscosity η of a monatomic gas is given by \eta = \frac{5}{16 \sigma^2} \sqrt{\frac{\pi m k_B T}{\Omega^{(2,2)}}}, where m is the molecular mass, k_B is Boltzmann's constant, T is the temperature, σ is the characteristic collision diameter from the intermolecular potential (e.g., Lennard-Jones), and Ω^{(2,2)} is the collision integral for viscosity, which depends on the reduced temperature T^* = k_B T / ε (with ε the potential well depth).[73] This formula arises from integrating the linearized Boltzmann collision operator over the perturbation to the Maxwellian distribution. For simple potentials like hard spheres (where Ω^{(2,2)} = 1), the viscosity simplifies to η = (5/16 σ²) √(π m k_B T), independent of density in the dilute limit. The temperature dependence of viscosity follows η ∝ T^s, where the exponent s varies with the intermolecular potential: s = 0.5 for hard spheres, approaching s = 1 for long-range inverse-power potentials, and typically s ≈ 0.6–0.8 for realistic Lennard-Jones potentials used in noble gases. This arises because the collision integral Ω^{(2,2)} decreases with increasing T^*, softening the effective repulsion at higher temperatures.[73] The theory has been validated extensively for noble gases like helium, neon, argon, and krypton, where predictions using Lennard-Jones parameters match experimental viscosities to within 1–2% over wide temperature ranges at low densities. For example, for argon near room temperature, the computed η ≈ 22.7 μPa·s aligns closely with measured values, confirming the accuracy of the first-order expansion for monatomic systems. Extensions to polyatomic gases incorporate internal degrees of freedom via the Wang Chang–Uhlenbeck formalism, which modifies the Boltzmann equation to include rotational and vibrational energy distributions, enabling predictions of both shear and bulk viscosities while retaining the Chapman–Enskog perturbation structure. This approach has been applied successfully to diatomic gases like nitrogen and oxygen, adjusting collision integrals for anisotropic potentials.[74]Models for Liquids and Mixtures
For pure liquids, the Eyring theory provides a fundamental activated-rate approach to viscosity, treating flow as a thermally activated process where molecules overcome an energy barrier to shear. The theory posits that the shear viscosity \eta is given by \eta = \frac{h N_A}{V} e^{\Delta G / RT}, where h is Planck's constant, N_A is Avogadro's number, V is the molar volume, \Delta G is the Gibbs free energy of activation, R is the gas constant, and T is the absolute temperature. This model, derived from absolute reaction rate theory, successfully correlates viscosity with temperature dependence in simple liquids like water and hydrocarbons, emphasizing the role of molecular rearrangements in dense fluids. Building on Eyring's framework, the significant structure theory of liquids models the liquid state as a quasi-lattice with a fraction of gas-like and solid-like degrees of freedom, enabling predictions of transport properties including viscosity. In this approach, viscosity arises from the balance between vibrational (solid-like) and translational (gas-like) contributions, with the theory expressing \eta through partition functions that partition the liquid's configurational space.[75] Developed in the 1960s, it has been applied to compute viscosities of metals and organic liquids, offering insights into how structural disorder influences flow resistance beyond simple activation models.[76] For mixtures, particularly liquid blends, the Arrhenius mixing rule approximates the logarithm of the mixture viscosity as a weighted sum of the pure-component logarithms: \ln \eta_{\text{mix}} = \sum x_i \ln \eta_i, where x_i are mole fractions and \eta_i are pure viscosities. This empirical relation, rooted in reaction rate theory, performs well for miscible non-polar liquids like hydrocarbon blends at moderate concentrations, capturing the exponential temperature sensitivity of flow.[77] In contrast, for gas mixtures, the Wilke equation extends kinetic theory by weighting pure-component viscosities with collision factors: \eta_{\text{mix}} = \sum y_i \frac{\eta_i}{\sum y_j \phi_{ij}}, where y_i are mole fractions and \phi_{ij} account for molecular interactions; though derived for dilute gases, it informs hybrid models for vapor-liquid mixtures.[78] In suspensions of rigid particles in liquids, the Einstein formula predicts an increase in viscosity due to hydrodynamic interactions: \eta = \eta_0 (1 + 2.5 \phi), where \eta_0 is the solvent viscosity and \phi is the volume fraction of spheres. Valid for dilute regimes up to approximately 5% solids by volume, this relation highlights the rotational and translational perturbations caused by suspended particles, as derived from low-Reynolds-number hydrodynamics. Extensions beyond this limit incorporate higher-order terms for denser suspensions, but the linear form establishes the baseline scaling for colloidal systems like paints and slurries.[79] For electrolyte solutions, the Jones-Dole equation models relative viscosity as \eta_r = 1 + A c^{1/2} + B c, where c is concentration, A reflects ion-ion interactions (often negligible in dilute limits), and B the ion-solvent solvation effects. Proposed in 1929, it quantifies how electrolytes like NaCl increase viscosity through hydration shells, with positive B values indicating structure-making ions and negative for structure-breaking ones, applicable up to about 1 M. Recent computational advances, such as quantitative structure-activity relationship (QSAR) models for polymer solutions, leverage machine learning on molecular descriptors to predict viscosities; for instance, physics-informed neural networks combining molecular dynamics simulations with experimental data achieve high accuracy for perfluoropolyether lubricants and other polymers, addressing gaps in traditional empirical fits for complex solvents.[80][81] These steady-state models assume time-independent Newtonian behavior and have limitations in capturing transient effects, such as thixotropy in non-Newtonian mixtures where viscosity evolves with shear history.[82]Examples and Data
Water
Water serves as a benchmark Newtonian liquid in viscosity studies, exhibiting ideal linear response to shear stress without time-dependent effects, making it a standard reference for calibration and theoretical comparisons. Its viscosity properties are extensively documented, providing a foundation for understanding fluid behavior in aqueous systems. The dynamic viscosity (η) of water displays a pronounced temperature dependence, reaching a maximum of approximately 1.79 mPa·s at 0°C, where strengthened hydrogen bonding creates a more rigid molecular network that resists flow.[83] As temperature rises, thermal agitation weakens these bonds, reducing η to 1.002 mPa·s at 20°C and further to 0.282 mPa·s at 100°C.[83] This behavior includes an anomalous maximum near 0°C, arising from structural transitions in the liquid phase where cooling promotes the formation of transient, ordered clusters that enhance intermolecular cohesion beyond simple thermal expectations.[84] Kinematic viscosity (ν = η / ρ), which accounts for density (ρ), couples these effects and is particularly relevant for applications involving gravitational flow. Water's density peaks at 4°C (approximately 1000 kg/m³), leading to a nuanced temperature profile for ν; for example, it measures 1.787 × 10^{-6} m²/s at 0°C, 1.004 × 10^{-6} m²/s at 20°C, and 0.294 × 10^{-6} m²/s at 100°C, reflecting the interplay between decreasing η and varying ρ.[83] Under pressure, water's viscosity remains largely unaffected up to 100 MPa, with increases typically less than 1% at ambient temperatures, owing to its low compressibility that preserves the hydrogen-bonded structure.[85] This stability underscores water's role as a reliable medium in high-pressure engineering contexts.Air
Dry air serves as a standard reference for gas-phase viscosity due to its prevalence in atmospheric and engineering applications. The dynamic viscosity of dry air at 20°C and 1 atm is approximately 18.1 μPa·s.[86] This value aligns well with predictions from the Chapman–Enskog theory, which derives transport properties from kinetic theory for dilute gases like air.[87] The dynamic viscosity of air increases with temperature, approximately following a T^{0.7} dependence as described by the Sutherland formula: \mu = \mu_0 \left( \frac{T}{T_0} \right)^{3/2} \frac{T_0 + S}{T + S}, where \mu_0 is the reference viscosity at temperature T_0, and S \approx 110.4 K is the Sutherland constant for air.[3] At standard temperature and pressure (STP, defined here as 20°C and 1 atm for atmospheric contexts), the kinematic viscosity \nu = \mu / \rho is approximately $15.1 \times 10^{-6} m²/s, where \rho \approx 1.20 kg/m³ is the density of dry air.[88] The presence of water vapor slightly increases the viscosity of air, with measurements showing an enhancement of up to a few percent at high relative humidities (e.g., 90%) and moderate temperatures (20–50°C), due to the higher viscosity of water vapor compared to dry air components.[89] This effect is generally small and often negligible for many applications unless humidity exceeds 50%.[90] In the Earth's atmosphere, air viscosity varies primarily with temperature under the U.S. Standard Atmosphere model, decreasing in the troposphere (up to ~11 km altitude) as temperature drops from 15°C at sea level to -56°C, then increasing in the stratosphere due to rising temperatures. Dynamic viscosity remains nearly independent of pressure (altitude-induced density changes), with values ranging from ~17 μPa·s at sea level to ~12 μPa·s at 11 km and back toward ~18 μPa·s at 20 km.[91]Other Common Substances
Beyond water and air, viscosity varies widely among other common substances, influenced by molecular structure, temperature, and shear conditions. For instance, everyday fluids like honey exhibit dynamic viscosities typically ranging from 2 to 10 Pa·s at room temperature, though values can reach up to 23 Pa·s depending on moisture content and floral origin, and honey displays non-Newtonian behavior where viscosity decreases under shear.[92][93] Engine oils, critical for lubrication, have viscosities of approximately 0.005 to 0.015 Pa·s at operating temperatures (around 100°C) for typical SAE grades, but this is highly temperature-sensitive, dropping significantly as heat increases to maintain flow in engines.[94] Biological fluids such as human blood show apparent viscosities of 3 to 4 mPa·s at physiological shear rates, with shear-thinning properties that reduce resistance during circulation.[29] Among gases, carbon dioxide at standard temperature and pressure (STP) has a viscosity of approximately 1.5 × 10^{-5} Pa·s, while helium exhibits a similar low value of about 2.0 × 10^{-5} Pa·s at 20°C with notably weak temperature dependence compared to polyatomic gases.[95][96] Industrial materials like polymer melts and molten glass demonstrate much higher viscosities, often in the range of 10^3 to 10^6 Pa·s during processing, reflecting their entangled molecular networks that impede flow.[97][98] Recent chemical advances in viscosity reducers for heavy oils, including multi-effect formulations, have achieved reductions of 50% to 90% at elevated temperatures (50–90°C), enhancing extraction efficiency in reservoirs.[99] The following table summarizes representative viscosity values for these substances under typical conditions, illustrating the broad spectrum from low-viscosity gases to highly viscous melts:| Substance | Typical Viscosity (Pa·s) | Conditions/Notes | Source |
|---|---|---|---|
| Honey | 2–10 | At 20–25°C; non-Newtonian, shear-thinning | [93] |
| Engine Oil (SAE grades) | 0.005–0.015 | At 100°C for SAE 10–50; strong temperature dependence | [94] |
| Human Blood | 0.003–0.004 | At physiological shear rates (~100 s^{-1}); shear-thinning | [29] |
| CO_2 Gas | 1.5 × 10^{-5} | At STP (0°C, 1 atm) | [95] |
| Helium Gas | 2.0 × 10^{-5} | At 20°C; minimal T dependence | [96] |
| Polymer Melts | 10^3–10^5 | At processing temps (150–250°C) | [97] |
| Molten Glass | 10^3–10^6 | At working temps (800–1400°C) | [98] |
| Heavy Oil (with reducers) | 50–90% reduction from baseline (~10^3–10^4) | At 50–90°C with chemical additives | [99] |