Dean-Stark Extraction

Dean-Stark extraction is a laboratory distillation method used to measure fluid saturations in core samples recovered from oil and gas reservoirs. By boiling a solvent through the sample and collecting the vaporized water in a calibrated trap, technicians can directly quantify water volume and calculate oil content by mass difference. The technique is a cornerstone of Routine Core Analysis (RCAL) and provides critical input data for reservoir characterization, reserve estimation, and field development planning.

Principle of Operation

The Dean-Stark apparatus consists of a flask, a reflux condenser, and a graduated side-arm trap positioned between them. The core plug is placed in a thimble inside the flask, which is filled with an appropriate solvent. When heated to boiling, solvent vapor rises through the sample, carrying water vapor upward into the condenser. Condensed water, being denser than the solvent, collects in the calibrated trap where its volume can be read directly. The condensed solvent, being lighter, overflows the trap and drips back down over the sample in a continuous wash cycle. This reflux action progressively strips residual oil from the pore space as the extraction proceeds, typically over 8 to 24 hours depending on sample permeability and oil viscosity.

Solvent Selection

Toluene is the most commonly used solvent and is well suited to sandstone samples containing light to medium crude oils. It has a boiling point of 111 degrees Celsius, sufficient to vaporize connate water while remaining relatively gentle on mineral grains. Xylene is preferred for heavier oils and tighter formations, as its higher boiling point (138 to 144 degrees Celsius) provides more aggressive extraction. A chloroform-methanol azeotrope is used for carbonate samples where acidic mineral dissolution is a concern with aromatic solvents. Solvent selection is guided by API RP 40 and ASTM D95 standards and is typically matched to crude oil gravity and core lithology reported in the well log.

Measurement Procedure and Calculations

The sample is first weighed in its preserved state to establish bulk wet mass. After extraction is complete, the trapped water volume is recorded directly from the graduated trap, giving water saturation volume (Vw). The sample is then dried in an oven at 105 degrees Celsius to constant weight. The mass difference between the wet and dry states, minus the water mass, is attributed to oil removed during extraction. Oil volume is calculated by dividing the oil mass by an assumed or measured oil density. With grain volume determined by helium porosimetry and bulk volume from dimensional measurements, water saturation (Sw) and oil saturation (So) are expressed as fractions of pore volume. Gas saturation is calculated as the remainder: Sg = 1 minus Sw minus So.

Applications in Core Analysis

Dean-Stark saturation data forms the foundation of RCAL workflows alongside porosity, permeability, and grain density measurements. The water saturation values inform log calibration, helping petrophysicists tie resistivity log responses to actual fluid content in the reservoir interval. When combined with capillary pressure data from Special Core Analysis (SCAL), Dean-Stark results help define irreducible water saturation (Swirr) and residual oil saturation (Sor), which are essential inputs for relative permeability modeling and enhanced oil recovery design. The method is also applied to preserved sidewall cores and conventional whole cores from exploration and appraisal wells where saturation state is a primary deliverable.

Accuracy Limitations and Sources of Error

Several factors can compromise Dean-Stark accuracy. Clay-bound water and capillary-bound water may not fully vaporize under standard conditions, causing underestimation of total water saturation. Volatile light hydrocarbons (C1 to C5 fractions) are partially lost during boiling, leading to underestimation of oil saturation in gas-condensate or light oil reservoirs. Carbonate samples are susceptible to partial dissolution by toluene or xylene, which inflates apparent porosity and distorts saturation calculations. Sample invasion by drilling mud filtrate before retrieval is another common source of error: water-based mud filtrate displaces original connate water and displaces some oil, so measured saturations may not represent in-situ reservoir conditions. Preservation of cores in aluminum foil and wax immediately after recovery reduces filtrate invasion effects and is required for reliable saturation measurements.

Standards and Regulatory Context

The method is standardized under ASTM D95 (Standard Test Method for Water in Petroleum Products and Bituminous Materials by Distillation) and API RP 40 (Recommended Practices for Core Analysis). Most national energy regulators, including the Alberta Energy Regulator and the U.S. Bureau of Land Management, accept Dean-Stark saturation data as part of core analysis reports submitted with well license applications and reservoir description filings. Laboratories performing Dean-Stark extraction for regulatory submissions are expected to document solvent type, extraction duration, trap readings, and drying conditions in their analytical reports.

Key Takeaways

• Dean-Stark extraction measures water volume directly from a calibrated trap and derives oil saturation by mass difference after drying, making it one of the most reliable direct saturation methods available in RCAL workflows.
• Solvent choice matters: toluene suits light oil sandstones, xylene handles heavier crudes, and chloroform-methanol azeotrope is preferred for carbonates to avoid mineral dissolution.
• Clay-bound water, volatile hydrocarbon loss, and mud filtrate invasion are the primary sources of error that must be accounted for when interpreting Dean-Stark results for log calibration or reserve calculations.
• Results are governed by ASTM D95 and API RP 40 and feed directly into SCAL programs, petrophysical log calibration, and reservoir simulation model initialization.