Distillation Extraction

Key Takeaways

• Dean-Stark water content measurement is considered the most reliable direct method for connate water saturation determination, but requires careful handling from the wellbore to the laboratory — the accuracy of the distillation extraction measurement depends critically on preserving the in-situ fluid distribution from the moment the core is cut until the analysis is performed; upon recovery, dissolved gas in the core fluids exsolves as pressure drops from reservoir to surface conditions, and light hydrocarbons (C1-C5) evaporate from the cut face of the core unless it is sealed immediately; refrigeration of core samples during transport slows evaporation and prevents bacterial degradation of oil; for highest-accuracy fluid saturation measurements, preserved core (sealed in aluminum foil and wax immediately at the wellsite and maintained at low temperature throughout transport and storage) is analyzed using a modified extraction technique that accounts for light hydrocarbon loss; routine core analysis on conventionally handled (unpreserved) samples may underestimate oil saturation by 20-40% due to evaporative loss of light ends between coring and extraction.
• The Dean-Stark extraction provides the reference point against which logging-derived saturations must be calibrated — wireline log interpretation (using Archie's equation or its shaly-sand modifications) calculates water saturation from measured formation resistivity, but this calculation requires accurate values of cementation exponent (m), saturation exponent (n), and formation water resistivity (Rw) that can only be determined reliably from core measurements; the Dean-Stark extracted water saturation from preserved core provides a direct measurement of Sw at the specific depth where the log tool made its resistivity measurement, allowing the log analyst to calibrate the n and m parameters by finding the values that make the Archie equation match the core-measured Sw; without core calibration, log-derived saturations may be systematically biased — overestimating Sw in carbonate reservoirs with complex pore geometry (where m can deviate significantly from the sandstone default of 2.0), or underestimating Sw in clay-rich sands (where the parallel conductance of clay minerals causes Archie's equation to underestimate true Sw without shaly-sand corrections).
• Solvent selection for distillation extraction must be matched to the oil type to ensure complete oil removal from pore surfaces — light paraffinic oils dissolve readily in common extraction solvents like toluene or naphtha; heavy asphaltic oils (high asphaltene content) may precipitate asphaltene particles when contacted with certain solvents (particularly aliphatic solvents like naphtha that cause asphaltene flocculation), leaving asphaltene deposits in the pore throats that are not removed by the extraction; this incomplete oil removal underestimates porosity (the pore volume appears smaller because asphaltene-clogged pores are not measured after drying) and overestimates residual oil saturation; for heavy oil and bitumen samples, chloroform-methanol azeotrope or other polar solvent mixtures provide better extraction completeness than single-component solvents; the selection of appropriate solvent systems for the specific crude oil type encountered in a well is part of the core analysis program design, and using the wrong solvent can produce systematically biased results that propagate into incorrect reserve estimates if the biased core data is used to calibrate log interpretations.
• Post-extraction porosity measurements using restored state or as-received methods complement the Dean-Stark fluid saturations to build a complete core analysis dataset — after Dean-Stark extraction and drying, the cleaned core plug is used for porosity and permeability measurements (gas expansion porosimetry, liquid permeability, Klinkenberg-corrected gas permeability) that quantify the pore volume and flow capacity of the formation; the combination of Dean-Stark water saturation with cleaned plug porosity and permeability from the same sample provides the input data for the pore-level calculations (irreducible water saturation, relative permeability, capillary pressure) that determine how the formation will produce once the well is on production; this integration of fluid saturation (from Dean-Stark) with pore structure data (from cleaned plug measurements) is the foundation of the full core analysis report that geoscientists and reservoir engineers use to build the reservoir model and predict field performance.
• Retort distillation is an alternative extraction method for rapid field measurement but trades accuracy for speed — the retort method places a core sample in a sealed retort chamber and heats it to approximately 650°C, vaporizing and condensing both the oil and water into calibrated receiving tubes for direct volume measurement; the retort is faster than Dean-Stark (minutes versus hours) and can be performed in a well site laboratory for real-time formation evaluation; however, retort results are subject to several systematic errors: water of crystallization released from clay minerals at high temperature is counted as pore water and overestimates water saturation in clay-rich formations; oil is cracked at retort temperatures and the condensed cracked hydrocarbons have lower volume than the original oil (requiring a correction factor that varies with oil gravity); and the high temperature alters clay structure and pore geometry, making the cleaned sample unsuitable for subsequent permeability or capillary pressure measurements; retort data is useful for real-time formation evaluation decisions at the wellsite but should be replaced by Dean-Stark analysis on the same samples for any detailed reservoir characterization or log calibration work.