Dispersion
Dispersion in the oilfield context refers to the spatial separation and mixing of fluid components as they flow through porous rock — a process by which an initially sharp boundary between two fluids (such as the flood front between injected water and reservoir oil in a waterflood, or the front between a chemical EOR slug and the chasing water) becomes progressively more diffuse and spread out as the fluids travel through the pore network, driven by the combination of molecular diffusion (the random Brownian-motion-driven migration of dissolved species from high-concentration to low-concentration regions), mechanical dispersion (velocity variations between flow paths of different sizes and geometries that cause faster-moving fluid fingers to overtake slower portions), and chromatographic effects (differential adsorption of EOR formulation components onto mineral surfaces that retards some components relative to others); hydrodynamic dispersion quantifies the combined mechanical and diffusive mixing as a single transport coefficient with units of length squared per time (m^2/s), with the ratio of convective transport to diffusive transport described by the Péclet number (Pe = v × L / D, where v is velocity, L is characteristic length, and D is the molecular diffusion coefficient), and dispersion determines how efficiently an EOR slug or water flood maintains its composition and concentration as it travels through the reservoir, directly affecting the ultimate recovery efficiency of the process.
Key Takeaways
- Hydrodynamic dispersion in porous media combines two fundamentally different mixing mechanisms into a single effective transport coefficient — molecular diffusion (a random walk process described by Fick's Laws with diffusion coefficient D of approximately 10^-9 m^2/s for dissolved species in water at reservoir temperatures) and mechanical dispersion (the velocity spread caused by pore-scale flow path heterogeneity, proportional to the flow velocity with dispersion coefficient alpha × v, where alpha is the dispersivity in meters); at low flow rates typical of reservoir flow (pore velocity below 10^-5 m/s), molecular diffusion dominates (Pe less than 1) and the two processes are comparable in magnitude; at higher flow rates, mechanical dispersion dominates (Pe much greater than 1) and increases linearly with flow velocity while molecular diffusion remains constant; the dispersivity (alpha) characterizes the length scale of pore-scale heterogeneity responsible for mechanical dispersion and typically ranges from millimeters in a homogeneous sandstone to centimeters in a heterogeneous carbonate, increasing to meters or tens of meters at the field scale where macroscopic permeability heterogeneity (layering, fractures, high-permeability streaks) contributes to large-scale dispersion.
- Chromatographic separation in EOR processes occurs when the multiple components of a chemical formulation (surfactant, polymer, alkali, cosolvents) travel through the porous rock at different velocities because of differential adsorption onto mineral surfaces — surfactant molecules adsorb strongly onto clay minerals and carbonate surfaces, reducing the surfactant concentration in the flowing phase and retarding the surfactant front relative to the water front; polymer molecules may be excluded from small pores (inaccessible pore volume effect) causing the polymer to advance faster than would be predicted from the pore volume injection count alone; alkali components may react with the formation to form new minerals or dissolve existing carbonate, consuming alkali and retarding the alkali front; the combination of these differential retardation effects separates the originally co-injected ASP (alkali-surfactant-polymer) slug into three fronts arriving at different times at the production well, reducing the simultaneous presence of all three components that is needed to achieve the target ultra-low interfacial tension required for enhanced oil mobilization.
- Dispersivity scale dependence is one of the most practically important and technically debated aspects of dispersion in reservoir engineering — laboratory measurements of dispersivity on 10-to-30-centimeter core plugs give values of 0.001 to 0.01 meters, while field-scale tracer tests give apparent dispersivities of 1 to 100 meters for the same reservoir, a scale factor of 100 to 10,000 times; this scale dependence arises because dispersivity measured at the core scale captures only pore-scale heterogeneity, while field-scale apparent dispersivity reflects the cumulative effect of all heterogeneity scales from pore-scale to bed-scale to interwell-scale, including permeability layering, geological facies variations, and fracture networks that cause bypassing at much larger scales than captured by a core plug; the practical consequence is that dispersion parameters measured in the laboratory cannot be directly scaled to field conditions, and field-scale dispersion must be estimated from inter-well tracer tests or tuned through history matching of field-scale EOR or waterflood production responses.
- Tracer dispersion in inter-well tracer tests directly measures the field-scale hydrodynamic dispersion between injectors and producers — when a tracer slug is injected into an injector well, the tracer arrives at the surrounding production wells as a spread-out pulse rather than a sharp peak, with the width and shape of the arrival pulse determined by the dispersion characteristics of the flow path between the two wells; the first-moment (mean arrival time) of the tracer pulse provides the average interstitial velocity and therefore the swept pore volume between the wells, while the second moment (variance of the arrival time distribution) provides the field-scale dispersivity that quantifies how broadly the chemical EOR slug will spread as it travels through the same reservoir interval; the ratio of the tracer peak arrival time to the tracer pulse width provides the Péclet number for the interwell flow path, directly measuring whether the transport is dispersion-dominated (Pe below 10, slug spreads significantly before reaching producer) or convection-dominated (Pe above 100, slug arrives relatively compact at the producer).
- Compositional reservoir simulation accounts for dispersion effects on EOR process efficiency through the numerical dispersion (an artifact of discretizing the flow equations on a finite grid that causes artificial mixing between grid cells, particularly problematic for coarse grids) versus physical dispersion (the real mixing due to pore-scale and field-scale heterogeneity); coarse-grid compositional simulations of ASP flooding or CO2 miscible displacement routinely generate numerical dispersion equivalent to dispersivities of 10 to 50 meters (proportional to the grid cell size, which is alpha_numerical = delta_x / 2 for first-order upwind discretization), which masks the physical dispersion that would occur at the same scale; accurate EOR simulation requires either fine enough grids to minimize numerical dispersion (typically requiring grid cells of 1 to 5 meters in the direction of flow), or the use of higher-order discretization schemes that reduce numerical dispersion at the same grid resolution, or explicit dispersion coefficient specification in the simulator that combines physical and numerical dispersion to match the known interwell tracer test dispersivity.
Fast Facts
The scale dependence of dispersivity in porous media — where dispersivity increases from millimeters at the laboratory scale to hundreds of meters at the field scale — is one of the most persistently challenging problems in reservoir engineering and groundwater hydrology. The Danish engineer Philip Scheidegger provided the first theoretical framework for mechanical dispersion in porous media in 1954, and the scale dependence was empirically documented by Gelhar et al. in a landmark 1992 review in Water Resources Research that compiled dispersivity measurements from over 200 field studies and showed a systematic increase of dispersivity with observation scale spanning 12 orders of magnitude in scale. This review fundamentally changed how reservoir engineers think about dispersion — not as a fixed property of the rock but as a scale-dependent effective parameter that reflects the statistical nature of subsurface heterogeneity and cannot be reliably extrapolated from core to field scale without field-scale calibration data from inter-well tracer tests or production history matching.
What Is Dispersion?
Inject a slug of tracer into a well, then watch for it at the production wells. In a perfect, homogeneous porous medium with uniform velocity everywhere, the tracer would travel exactly one pore volume through the rock and arrive all at once at the producer — a perfect sharp breakthrough. In the real subsurface, it arrives as a spread-out plume: some particles travel fast through high-permeability paths, others wander through tortuous low-permeability paths, and some get slowed by diffusing into and out of dead-end pores. The result is a dispersed arrival curve — a broad, rounded breakthrough profile that tells you about the distribution of travel times through the formation.
This spreading is dispersion, and it has real economic consequences for enhanced oil recovery. When a chemical EOR slug disperses as it travels through the reservoir, it arrives at the producing wells diluted and separated rather than at the designed concentration. The ultra-low interfacial tension that requires a precise surfactant concentration may never be achieved if the surfactant slug disperses below the critical concentration before it reaches the remaining oil. Understanding and controlling dispersion — through tracer test measurement, simulation matching, and injection strategy optimization — is part of the engineering discipline that determines whether a chemical EOR project achieves its recovery target or falls short because the chemicals never reach the oil at the designed conditions.
Dispersion in EOR and Waterflood Processes
Miscible displacement efficiency in CO2 and hydrocarbon gas injection projects is directly affected by dispersion — when CO2 is injected into a reservoir containing oil that is miscible with CO2 at first-contact or multiple-contact conditions, the theoretical displacement efficiency at microscopic scale is 100% (CO2 completely displaces oil from contacted pore space), but macroscopic efficiency is reduced by channeling through high-permeability intervals and by mixing (dispersion) at the CO2-oil interface that creates a transition zone of CO2-oil mixture extending ahead of and behind the nominal flood front; the transition zone grows in proportion to the square root of dispersion distance (the "root-DL" spreading), diluting the CO2 concentration behind the front and contaminating the oil ahead of it with dissolved CO2, reducing the density and viscosity of the transition zone fluid in a way that may improve or worsen the mobility ratio depending on the specific reservoir and fluid system; polymer flooding reduces viscous fingering (a form of convective mixing driven by adverse mobility ratio between the injected water and the displaced oil) by increasing the viscosity of the injected phase to approach the reservoir oil viscosity, effectively reducing the mechanical dispersion component that drives finger development at the flood front.
Breakthrough curves in chemical injection pilots are interpreted using the convection-dispersion equation (CDE, the 1D transport equation C/t = D × C_xx - v × C_x, where C is concentration, D is dispersion coefficient, v is average velocity, and subscripts denote derivatives) to extract the interstitial velocity and dispersivity that characterize the pilot's transport behavior — fitting the measured breakthrough curve shape to the CDE solution gives the mean travel time (first moment) and the dispersivity (proportional to the variance divided by the mean travel time), providing the transport parameters that can then be used to predict tracer and chemical front behavior at larger scales and to estimate the required slug size for a full-field EOR project that accounts for the concentration reduction from dispersion during travel from injector to producer.
Dispersion Across International Jurisdictions
Canada (AER / WCSB): WCSB waterflood operations in Cardium, Viking, and Pembina fields have used inter-well tracer tests since the 1970s to characterize dispersion and flow channeling that affect waterflood sweep efficiency and EOR project design; AER's EOR scheme applications require technical assessments that include characterization of reservoir heterogeneity and connectivity, which in practice involves tracer test results that document the distribution of interwell travel times and apparent dispersivities in the production pattern; Canadian unconventional oil (Athabasca oil sands SAGD operations) presents a different dispersion context — steam dispersion in the reservoir and thermal dispersion of heat within the SAGD chamber are governed by analogous transport equations, and the thermal dispersivity of the Athabasca sands (measured from temperature monitoring in SAGD observation wells) provides calibration data for reservoir simulation models that predict the growth of the steam chamber and the recovery efficiency of SAGD pairs.