It is widely accepted that fluid flow is an important aspect of prograde metamorphism. Supercritical fluids released by prograde metamorphic devolatilisation reactions are mobile, and potentially highly reactive. Fluid flow
is buoyancy-driven, and fluids may alter the chemistry, mineralogy, temperature and rheology of the rocks along their flow-path. Various theoretical models have been used to constrain fluid fluxes in 1-D based on cross-layer propagation of isotope or reaction fronts. An integral assumption of these models is that isotopic and reaction fronts are planar and parallel to layering. This assumption is challenged based on the 2-D morphologies of isotope and reaction fronts in metabasic sills which are delineated and investigated herein.
At Port Ellen, Islay, SW Scottish Highlands, a suite of pre-metamorphic metabasite sills is emplaced within
phyllites of the Dalradian Supergroup. The sequence
was folded into a recumbent NW-facing antiform and
subsequently metamorphosed in the greenschist facies
(T = 470±30C, P = 9±1 kbar).
During metamorphism, infiltration of a CO2-bearing hydrous fluid caused carbonation of the metabasite
assemblage: amphibole + epidote + albite ± chlorite, by the reaction:
3 amphibole + 2 epidote + 8 H2O + 10 CO2 =
3 chlorite + 10 calcite + 21 quartz
Reacted, carbonate-bearing, schistose sill margins are separated from the unreacted, carbonate-free, massive sill interior by reaction front zones within which amphibole
and calcite co-exist. Asymmetric propagation of these reaction front zones has been used to infer the direction and magnitude of fluid fluxes (Skelton et al., 1995). Propagation of reaction fronts was coupled with propagation of oxygen isotopic fronts. In general, d18O of the sill margins approaches that of the host phyllites (ª 12 ), whereas d18O of the sill interior approaches the pre-metamorphic value of 6-8 . Various 1-D and 2-D models exist to describe the coupling (or decoupling) of reaction and isotope fronts based on the combined influences of advection, diffusion and mechanical and chemical kinetic dispersion. The principle aim of this study was to test these models by investigating front geometries in 2-D.
d18O and reaction progress were measured at 0.5-1.5m intervals (0.1-0.5m, in critical regions) and from a 15 ¥ 47m grid. The 2-D geometries of front propagation from lower and upper sill margins are different.
Both isotope and reaction fronts which propagated
from the lower margin (with the inferred direction of fluid flow, Skelton et al., 1995) are significantly broadened, approximately planar and parallel to the sill margin. Within sampling resolution, they are coincident. However, broadening of the isotopic front is more extensive than broadening of the reaction front. This geometry can be modelled by almost pure cross-layer diffusion.
Isotope and reaction fronts which propagated from
the upper margin (opposing the inferred direction of fluid flow, Skelton et al., 1995) are highly irregular in 2-D. The geometry of the 'coupled front' is best described by 'nodes' of high d18O reacted metabasite. In 3-D, these may relate to 'rod-' or 'sheet-'like protrusions which emanate from the sill boundary. Where the 'coupled front' is coincident with the sill boundary, the isotope front is broadened by several metres, whereas the 'node' boundaries are defined by sharp (<5cm wide) and precisely coincident reaction and isotope fronts. Therefore advection was parallel to the node boundaries and across the sill margin. Evidently the nodes
were 'pipes' or 'cracks' along which fluid from the adjacent phyllites penetrated the upper margin of the sill. This geometry can be modelled by coupled advection and kinetic dispersion, where kinetic dispersion results from transverse diffusion outwards from inferred fluid 'pipes' or 'cracks'.
Skelton, A.D.L., Graham, C.M. & Bickle, M.J. J. Petrol. 36, 563-586 (1995).