Constant Composition Expansion (CCE/CME) Test

PVT Analysis of Reservoir Fluids › 1. CCE / CME — Constant Composition (Mass) Expansion

Reviewed by Coretest reservoir-fluids engineers · Updated 21 June 2026

The Constant Composition Expansion (CCE), historically also called Constant Mass Expansion (CME) or flash expansion, is the foundational depletion experiment in any black-oil, volatile-oil or gas-condensate PVT study. It is almost always the first test performed after recombination/validation of the sample, because it locates the saturation pressure that anchors every subsequent test (differential liberation, separator tests, CVD).

Purpose↑ Contents

CCE answers the central single-phase reservoir question: how does a reservoir fluid of fixed overall composition respond, volumetrically, as pressure is reduced isothermally at reservoir temperature, and at what pressure does the first bubble of gas (oil) or first drop of liquid (gas-condensate) appear. Concretely it delivers:

• The saturation pressure: bubble-point pressure \(p_b\) for oils, dew-point pressure \(p_d\) for condensates.
• The pressure-volume (P-V) relationship above and below saturation, which is the backbone for all later PVT computations and the primary data set used to tune an equation of state (EOS).
• The isothermal oil compressibility \(c_o\) in the single-phase (undersaturated) region, used in material balance, reservoir simulation and geomechanics.
• The relative volume curve, the normalized P-V relation against which a PR or SRK EOS is regressed.
• A qualitative fluid-type classification (black oil vs volatile oil vs gas-condensate) from the shape and slope discontinuity of the P-V curve.

A note on terminology and current practice: CCE and CME describe the same closed-cell, constant-composition experiment. For oils, CCE remains a mandatory, modern, EOS-relevant test. For gas-condensates the depletion behaviour relevant to a producing reservoir is captured instead by Constant Volume Depletion (CVD), so the CCE on a condensate is used essentially only to fix the dew point and the single-phase \(Z\), while CVD provides the produced-fluid and liquid-dropout data that simulators (PVTsim, PVTSOL, Eclipse E300, tNavigator) actually ingest for depletion. The "CME removes gas" reading found in some teaching material is non-standard: in proper usage the cell stays closed and total mass and composition are constant. Where gas is progressively withdrawn it is no longer a CCE/CME but a differential or constant-volume process.

Physical principle↑ Contents

Above the saturation pressure the fluid is a single liquid (oil) phase. Reducing pressure isothermally expands that liquid only slightly, governed by its isothermal compressibility:

$$c_o = -\frac{1}{V}\left(\frac{\partial V}{\partial p}\right)_T$$

where \(c_o\) is isothermal oil compressibility (\(\text{1/psi}\) or \(\text{1/bar}\)), \(V\) is the fluid volume in the cell, \(p\) is pressure, and the subscript \(T\) denotes constant temperature. Because the liquid is nearly incompressible, the single-phase P-V line is steep and close to linear (small \(dV\) per unit \(dp\)).

At the saturation pressure the lightest components begin to evolve a second phase. For an oil this is the bubble point \(p_b\): the first incremental gas phase appears. Below \(p_b\) the system is two-phase and total volume grows rapidly with each pressure drop because gas, being highly compressible and continuously evolving from solution, dominates the volumetric response. The P-V curve therefore exhibits a slope break at saturation:

$$\left(\frac{\partial V}{\partial p}\right)_{p > p_b} \ne \left(\frac{\partial V}{\partial p}\right)_{p < p_b}$$

This discontinuity in slope is the experimental signature of \(p_b\) for a black oil (sharp break). For volatile oils the break is softer; for gas-condensates a P-V plot alone is a poor dew-point indicator and the dew point is instead found by visual observation of the first liquid drop (windowed cell), which is why a clear sight cell or a constant-composition expansion with visual detection is used.

Phase equilibrium below saturation is described by the equality of component fugacities between phases, expressed through equilibrium ratios:

$$K_i = \frac{y_i}{x_i}$$

where \(K_i\) is the equilibrium ratio of component \(i\), \(y_i\) is its mole fraction in the gas (vapour) phase and \(x_i\) its mole fraction in the liquid phase. CCE does not measure \(K_i\) directly (no phase is removed for compositional analysis), but the P-V behaviour it produces constrains the K-value behaviour that the EOS must reproduce.

The volumetric behaviour of the evolved gas is described by the real-gas law through the deviation (compressibility) factor:

$$Z = \frac{pV}{nRT}$$

where \(Z\) is the gas compressibility (deviation) factor (dimensionless), \(p\) is absolute pressure, \(V\) is gas volume, \(n\) is moles of gas, \(R\) is the universal gas constant, and \(T\) is absolute temperature (\(\text{K}\)). \(Z < 1\) indicates the real gas is more compressible than ideal (typical at moderately high pressure and low temperature); \(Z > 1\) indicates it is less compressible than ideal (very high pressure, or temperatures above the critical); \(Z \approx 1\) at low pressure and high temperature where ideal behaviour is approached.

For data reduction the P-V curve is normalized to the saturation point as the relative volume:

$$V_{rel}(p) = \frac{V_{total}(p)}{V(p_{sat})}$$

where \(V_{rel}\) is the relative volume (dimensionless), \(V_{total}(p)\) is the total cell volume of the fluid (single-phase liquid above saturation, liquid plus gas below saturation) at pressure \(p\), and \(V(p_{sat})\) is the volume at the saturation pressure (the reference volume, \(V_{rel} = 1\) by definition at \(p_{sat}\)). Using \(V_{rel}\) rather than absolute volumes removes the dependence on the exact sample charge and is the form regressed against the EOS.

A complementary reduction for two-phase data below \(p_b\) is the Y-function, which linearizes the relative-volume data and is used to smooth and quality-check the bubble point:

$$Y = \frac{p_{sat} - p}{p\,(V_{rel} - 1)}$$

where \(p_{sat}\) is the saturation pressure, \(p\) is the current pressure, and \(V_{rel}\) is the relative volume at \(p\). A plot of \(Y\) versus \(p\) below \(p_b\) should be linear; departures flag bad points or an incorrect \(p_b\).

Apparatus and conditions↑ Contents

• Cell: a high-pressure, high-temperature variable-volume PVT cell. Two designs are common. A windowed (visual) cell with a movable piston allows direct observation of first bubble or first dew and accurate volume readout via piston position; this is preferred for modern, mercury-free work. A blind (non-visual) cell relies entirely on the slope-break method for saturation pressure. Mercury-based cells (mercury as the displacing/pressurizing fluid) are the historical standard but are being retired for health and environmental reasons in favour of piston/hydraulic mercury-free cells.
• Volume measurement: by calibrated piston displacement (positive-displacement pump) or, in legacy units, by mercury volume injected/withdrawn. Pressure is held and stepped by the same pump.
• Temperature: held constant at reservoir temperature \(T_{res}\) throughout (this is an isothermal experiment). The cell is in an air bath or jacketed thermostat. The cited cell orientation (for example tilted at 135 degrees, with 0 degrees defined as glass/window down) is a practical detail to position the sight glass so the first gas bubble collects against the window for unambiguous visual detection.
• Held constant: total composition and total mass (closed cell, nothing produced) and temperature. Varied: pressure (and therefore total volume).
• Initial condition: pressure is set well above the expected saturation pressure so the fluid is confidently single-phase.

Procedure↑ Contents

1. Charge the cell with a representative single-phase sample: either a validated downhole (bottomhole) sample or a recombined sample (separator oil plus separator gas recombined at the measured producing GOR). Bring the cell to reservoir temperature \(T_{res}\) and allow thermal equilibration.

2. Pressurize the fluid to a pressure clearly above the anticipated saturation pressure to ensure a homogeneous single liquid phase, and let it stabilize.

3. Record the reference single-phase volume and pressure. This is the first point on the P-V curve.

4. Reduce the pressure by a controlled increment (larger steps far above saturation, progressively smaller steps as saturation is approached). After each step, agitate/rock the cell to accelerate equilibrium, then wait for pressure and volume to stabilize.

5. At each stabilized stage record stage pressure \(p_{stage}\) and the corresponding fluid volume \(V(p_{stage})\). Continue through the single-phase region to map the steep, near-linear undersaturated branch.

6. As the slope of the P-V line begins to change (or, in a windowed cell, as the first gas bubble or first liquid drop is observed), reduce the pressure step size markedly to bracket the saturation pressure precisely.

7. Identify the saturation pressure: for oils, by the slope break in the P-V plot and/or first visible bubble; for condensates, by first visible liquid (dew). Record \(V(p_{sat})\) as the saturation reference volume.

8. Continue stepping pressure below saturation into the two-phase region. After each step, agitate to re-establish equilibrium between the evolved gas and the remaining liquid (do not remove any gas: the cell stays closed, composition constant), then record \(p_{stage}\) and total two-phase volume \(V_{total}(p_{stage})\).

9. Continue the expansion to the lowest pressure of interest (typically toward or below the expected first separator pressure / atmospheric, depending on cell limits), building the full two-phase branch.

10. Throughout, keep temperature constant and verify it at every stage; log any thermal excursion.

11. Reduce the data: normalize all volumes to \(V(p_{sat})\) to obtain the relative volume curve, compute \(c_o\) from the single-phase branch, compute \(Z\) for the evolved gas where applicable, and run the Y-function check below \(p_b\).

12. Report curves and tables and pass the relative-volume and saturation data to the EOS-tuning workflow.

Measured parameters↑ Contents

At each pressure stage the directly recorded quantities are:

• Stage pressure, \(p_{stage}\) (\(\text{psia}\) or \(\text{bar}\)), absolute.
• Total fluid volume at that stage, \(V(p_{stage})\) (\(\text{cm}^3\)), from piston position or mercury volume.
• Temperature, \(T\) (kept at \(T_{res}\)), verified each stage.
• Saturation pressure \(p_{sat}\) (\(p_b\) or \(p_d\)), recorded at the stage where the slope break and/or first phase change occurs, together with the reference volume \(V(p_{sat})\).
• Where a visual cell is used, the qualitative observation of first bubble / first dew and, optionally, the relative liquid volume in the two-phase region (fraction of cell occupied by liquid) by reading the gas-liquid interface position.

Calculated / derived parameters↑ Contents

• Relative volume:

$$V_{rel}(p) = \frac{V(p)}{V(p_{sat})}$$

Dimensionless; \(V_{rel} = 1\) at \(p_{sat}\); \(> 1\) below saturation and \(< 1\) (slightly) above it referenced the other way, but conventionally \(V_{rel} \ge 1\) across the expansion with the reference at \(p_{sat}\). This is the principal EOS-regression target.

• Isothermal oil compressibility (undersaturated region only, \(p_{stage} > p_b\)), in stepwise finite-difference form:

$$c_o = -\frac{V(p_{stage}) - V(p_{stage-1})}{\left[\,p_{stage} - p_{stage-1}\,\right]\,V(p_{stage})}$$

Equivalent to \(c_o = -\frac{1}{V}\left(\frac{\partial V}{\partial p}\right)_T\). Units \(\text{1/psi}\) or \(\text{1/bar}\). Because \(c_o\) varies mildly with pressure, it is normally reported as an average over the undersaturated interval, or as a value at a stated reference pressure.

• Bubble-point (saturation) pressure \(p_b\): located as the slope-break point of the P-V curve where \(\left(\frac{\partial V}{\partial p}\right)\) above \(p_b\) differs from \(\left(\frac{\partial V}{\partial p}\right)\) below \(p_b\), refined with the Y-function linearization below saturation.

• Gas deviation factor \(Z\) for the gas phase evolved below the bubble point (or for the single-phase gas of a condensate above the dew point):

$$Z = \frac{pV}{nRT}$$

In ratio form referenced to standard conditions:

$$Z_{stage} = \frac{V_{gas}(p_{stage}, T_{stage})\,\cdot\,p_{stage}\,\cdot\,(273 + T_{std})}{V_{gas}(p_{std}, T_{std})\,\cdot\,(273 + T_{stage})}$$

where \(V_{gas}(p_{stage}, T_{stage})\) is the gas volume measured at stage pressure and temperature, \(V_{gas}(p_{std}, T_{std})\) is the same gas expanded to standard conditions, \(T_{std}\) is the standard temperature and \(T_{stage}\) the stage temperature (\(\text{degrees C}\) in the \(+273\) conversion to kelvin), and \(p_{std}\) is standard pressure. \(Z\) values are central to gas formation volume factor \(B_g\) and gas compressibility \(c_g\) downstream.

• Y-function (quality-control / smoothing of two-phase data and \(p_b\)):

$$Y = \frac{p_{sat} - p}{p\,(V_{rel} - 1)}$$

Linear in \(p\) below \(p_b\) for a consistent data set.

Reporting and typical curves↑ Contents

The deliverables are:

• The P-V (or pressure vs relative-volume) curve, the primary plot. It is read as two branches meeting at the saturation pressure: a steep, near-linear undersaturated branch (small volume change per unit pressure, reflecting low \(c_o\)) and a much flatter, rapidly expanding two-phase branch below saturation. The intersection / slope break gives \(p_b\). A sharp, clean break indicates a black oil; a rounded transition with a large undersaturated-to-saturated volume change indicates a volatile oil; for a gas-condensate the P-V curve alone is ambiguous and the dew point comes from visual liquid-drop observation.
• The relative-volume table and curve, \(V_{rel}\) vs \(p\), used directly for EOS tuning.
• The Y-function plot below \(p_b\) (should be linear), used to validate \(p_b\) and discard outliers.
• Tabulated stage data: \(p_{stage}\), \(V(p_{stage})\), \(V_{rel}\), and the derived \(c_o\) over the undersaturated interval and \(Z\) where computed.
• A summary line giving \(p_b\) (or \(p_d\)) at \(T_{res}\), and \(c_o\) at a reference pressure.

These outputs feed PR/SRK EOS regression (PVTsim, PVTSOL) and reservoir simulators (Eclipse, tNavigator) as the single-phase volumetric and saturation constraints.

Practical notes and pitfalls↑ Contents

• Sample quality dominates everything. A non-representative or contaminated sample (lost light ends, water cut, OBM filtrate contamination, or wrong recombination GOR) yields a wrong \(p_b\). Always validate the sample (opening pressure vs bubble-point check, compositional consistency) before CCE.
• Equilibration time: below the bubble point the evolved gas must fully equilibrate with the liquid at each step. Rock/agitate the cell and wait for both pressure and volume to be truly stable; reading too early gives a falsely low expansion volume and distorts the two-phase branch. Near the bubble point use small pressure increments and longer hold times.
• Saturation-pressure detection: rely on the slope-break method for oils, but corroborate with visual observation when a windowed cell is available. For gas-condensates do not trust a P-V break for the dew point; use direct visual detection of first liquid, otherwise the dew point can be missed entirely (retrograde behaviour).
• Isothermal control: any temperature drift changes volumes and corrupts \(c_o\) and \(p_b\). Verify \(T\) at every stage; this is an isothermal test by definition.
• Dead volume and calibration: account for cell dead volume, valve/line volumes and piston-position calibration. Uncorrected dead volume biases absolute volumes and therefore \(c_o\); normalizing to \(V(p_{sat})\) (relative volume) mitigates but does not eliminate systematic errors.
• Mercury-free operation: legacy mercury cells give excellent volume resolution but carry health/environmental hazards; modern practice favours piston/hydraulic mercury-free cells. When migrating, re-verify volume calibration against a known standard.
• Step-size strategy: large steps far above saturation waste no information, but the region within a few hundred psi of \(p_b\) must be sampled finely to resolve the slope break; coarse stepping there is the most common cause of an imprecise bubble point.
• Closed-cell discipline: in a true CCE/CME nothing is withdrawn, so composition and mass are constant. If gas is removed at any stage the experiment is no longer CCE: do not blend differential-liberation or CVD steps into the same run, or the relative-volume curve and \(p_b\) interpretation become invalid.
• Use the Y-function to catch bad two-phase points before they are passed to EOS regression; a single off-line point in \(V_{rel}\) can pull a tuned EOS away from the true saturation pressure.