(204f) A New Experimental Methodology for the Study of Hydrocarbon Phase Behavior Under Confinement

Introduction

Advances in
drilling, hydraulic fracturing and geophysical methods have led to a massive
increase in hydrocarbon production from shale and tight reservoirs. These
reservoirs are characterized by their low permeabilities and small pores, as a
substantial fraction of the porosity resides in pores with characteristic
dimension of between 1 and 100 nm. Towards the lower end of this range, the
pore dimensions are comparable to the size of the hydrocarbon molecules and
confinement phenomena become evident. Broadly, confinement phenomena include
both capillary condensation and size exclusion effects. Well-developed
experimental methodologies exist for measuring excess adsorption of pure
substances on porous media but these approaches are difficult to apply to mixtures
at reservoir conditions. In this work we introduce a new methodology to study
confinement phenomena on the phase behavior of binary hydrocarbon mixtures
under reservoir conditions. We present results for two binary systems: (methane
+ pentane) and (methane + decane).

Experimental

The technique is
based on the non-visual static-synthetic method for bubble-point determination,
augmented by a material balance criterion. This approach allows one to
determine an average pore-filling density and an average pore-fluid composition
under the condition in which the fluid-saturated porous medium coexists with a
bulk fluid phase at its bubble point. The heart of the apparatus was a 10 mL
stainless-steel pressure vessel sealed with a metallic o-ring. The volume of
this vessel was determined as a function of temperature and pressure by means
of careful calibrations. In an experiment without porous media present, the
vessel would first be charged with a known mass of methane, determined
gravimetrically. A precision syringe pump would then be used to inject a liquid
hydrocarbon slowly while the vessel was maintained at a constant temperature. A
magnetic stirrer was used to promote phase equilibrium, and both injected
volume of hydrocarbon liquid and pressure were recorded continuously until a
pressure well above the bubble point was reached. A plot of the pressure
against either the injected volume or the overall mole fraction of one
component was then analyzed to determine the bubble point pressure and
composition as illustrated in Fig. 1.

An experiment
with porous media was conducted in a similar way except that the vessel was
first partially filled with the desiccated medium. Porous media were characterized
by nitrogen adsorption at 77 K and helium pycnometry, from which the pore-volume
distribution and skeletal density were obtained. With a porous medium present
in the vessel and a bubble-point condition existing in the bulk phase, the
following quantities were determined:

·        
mass of porous
medium

·        
skeletal volume
of the porous medium

·        
total pore volume

·        
vessel volume

·        
mass of methane

·        
mass of
hydrocarbon liquid

·        
temperature

·        
pressure

Fig. 1 Pressure p as a function of pentane mole fraction x
measured in a blank run at T = 350 K.

Furthermore, from
blank experiments without porous media the density and composition of the bulk
phase at its bubble point were determined as a function of pressure at given
temperature: see Fig. 2.

Fig. 2 Bubble-point composition and density of the (methane + pentane) at T
= 350 K.

Treating the pore
space as a separate phase, it was then possible to determine the mean density
and composition of the fluid within the pores. Since the pores accounted for
only a small fraction of the total system volume, the results were quite sensitive
to small errors in the measured quantities and, to assess this, a rigorous
Monte Carlo uncertainty analysis was carried out. 

Results

Experiments were
carried out with the fluid systems methane (1) + pentane (2) and methane (1) +
decane (2) and with either SBA-15 or carbon coated SBA-15. These structured mesoporous
media had pore diameters of about 8 nm and served as proxies for shale of low
and high carbon content. Measurements were carried out at temperatures between 320
K and 400 K and the pore selectivity S₂₁ with respect to the heavy hydrocarbon was determined
as follows:

                                                         ,                                                    (1)

where, p indicates pore and b bulk. Fig. 3 shows the results for methane + decane + SBA-15 at T = 350 K. The data show that the pore fluid composition is
enriched in decane and this effect is stronger for carbon-coated SBA-15 at
higher pressures. Similar  observations were made at other states that were
investigated in this fluid system. For methane + pentane + SBA-15,
selectivities were generally closer to unity.

Fig. 3 Selectivities for decane in the methane (1) + decane (2) + SBA-15
system at T = 350 K: left, initial filling pressure 7 MPa (p_b
≈ 17 MPa); right, initial filling pressure 9 MPa (p_b
≈ 21 MPa). Error bars indicate ± one standard deviation.

Conclusion

The experimental technique has sufficient
resolution to reveal some interesting behavior in the type of mesoporous media
investigated. The results of this study suggest that carbon-rich mesoporous
media preferentially adsorb longer-chain hydrocarbons under reservoir
conditions.