The problem of secondary migration is as old as the petroleum industry itself, yet no consensus has been reached on the mechanism. Here, I review existing mechanisms and propose a new one.
From the earliest days, the favored mechanism has been migration of petroleum in its own, separate phase. This hypothesis immediately presented problems: droplets have inadequate buoyancy to overcome capillary entry pressures of pore throats, so that they have to coalesce (how? a Catch-22) into long, continuous phase segments called slugs or stringers to overcome these entry pressures. But, there was a lack of supporting observations, such as:
- Petroleum stringers of the necessary length (10s of meters) and thickness (meters);
- Stringers waiting for take-off on top of source rocks or at permeability boundaries;
- Residual petroleum saturations (in the wake of stringers), according to capillarity theory;
- Differences in the ease or migration for different petroleum types, on account of their
widely different capillary entry pressures and viscosities.
The opposite end-member model of migration in water solution was found to be inadequate, because of the low solubilities and diffusion rates of petroleum. Flowing water had to be
invoked, but compactional and artesian flows are generally too slow to account for the volumes
of migrated petroleum. Also, wholesale convection of pore water in carrier beds is usually denied
by stratified water chemistry. The next step was to call upon migration of petroleum in water
in higher concentrations in the form of colloids or emulsions, but the problem was the large
soap-hydrocarbon ratios that seem to be required. The attractive feature of this hypothesis was
that microdroplets could pass through pore-throats without capillary impediment, and this was supported by experiments half a century ago.
In spite of the difficulties posed by migration of petroleum in its own, continuous phase, this
hypothesis became regarded as “well-established” by the mainstream petroleum media in the late
1970’s and 1980’s, (e.g. Berg, 1975; Schowalter, 1979; England et al., 1987). Yet, there were
actually no new data to support this claim! The only problem that proponents of this hypothesis
seemed to have was its use of the computationally slow Darcy’s law. So, they introduced faster
computational approaches, such as invasion percolation that make even less well-supported
assumptions, e.g., treating viscous forces as negligible!
In view of the problems of petroleum migration in either water solution, or in continuous separate phase, one finally returns to the metastable region of colloidal dispersions. Solutes tend to cluster in supersaturated, metastable solutions, and this should be particularly true for hydrophobic petroleum with its large interfacial tension with water. These colloidal clusters are
smaller than the pore throats of typical carrier beds, so they are not subject to capillary
resistance and migrate at velocities proportional to their radius r squared (according to Stoke’s Law).
Implications of the proposed hypothesis:
1) Cluster formation occurs in the metastable region between solution and exsolution and does not require solubilizers or emulsifiers.
2) There are no capillary constraints. Thus, the ratio of viscous to capillary forces (the Capillary Number) is essentially infinite.
3) The Stokesian flux of petroleum clusters Jsc, through pore throats that are large enough to
allow their passage without capillary restraint, can be compared with the separate phase Darcy flux Jdo through the same rock. To a first approximation, the ratio of these fluxes is
proportional to the ratio of the kinematic viscosities of petroleum and water: Water is generally less viscous than petroleum and the flux of petroleum in clusters is (much) faster than the Darcy flow of a continuous, separate phase.
4) The mechanism works similarly for all types of petroleum, because the constraining viscosity is that of water rather than highly variable petroleum.
5) The ranges of velocity and flux of petroleum are very great. The radii r of clusters and
microdroplets varies by about three orders of magnitude, so the velocities vary by about
six orders of magnitude (~ cm/yr to ~10 km/yr!)
6) The mechanism is self-adjusting. If the migration of clusters or droplets lags behind the
petroleum influxes anywhere in the migration network, the supersaturation of petroleum in the pore water increases. This causes the clusters to enlarge and speed up (with the square of their size), and v.v.
7) The mass flow rates are balanced throughout the migration system, so petroleum does not
accumulate anywhere except in the trap; mass continuity is satisfied automatically.
8) Brownian motion of clusters plays an important role, by (a) assisting their movement through pore throats and preventing their sticking to pore walls, and (b) creating a zone of supersaturation of clusters (~ 1 m thick) at the top of carrier beds (aided by gravity).
9) In spite of the all-important role of clusters and microdroplets in this mechanism, their
volume fractions in the pores are relatively small (generally ~ 0.001 or less).
10) Losses of petroleum into water solution are very small compared with the total quantities
of petroleum migrated, so the mechanism is very highly efficient (~100%): virtually no residual petroleum remains remains along migration pathways.
DATA: 6 febbraio 2013, ore 9.30-12.30 presso l'Aula Consiglio VII Piano Dipartimento di Matematica Politecnico di Milano
contatto:
